food science research papers

• Journal home
• Advance online publication
• About the journal
• J-STAGE home
• Food Science and Technology Re ...
• Development of β-conglycinin quantitative method in soybean foods using capillary electrophoresis-based immunoassay Kazuhiro Fujita, Megumi Suzuki, Kazushi Mizukoshi, Yushi Takahashi, Toshiaki Yokozeki, Izumi Yoshida, Mari Maeda-Yamamoto
• Technique for assessing the astringency of persimmon fruit by measuring the liposome aggregation Kota Kera, Shohei Makino, Risako Takeda, Aoi Shimeno, Masaya Hojo, Sadahiro Hamasaki, Akihito Endo, Masumi Iijima, Tsutomu Nakayama
• Changes in the flow properties of potato starch supplemented with super-fine eggshell powder Daisuke Nei, Tatsuki Kamata, Yasumasa Ando
• The clean taste of coffee depends on the sodium content of its extraction water Koji Imamura

Effects of Cooking Conditions on the Relationships Among Oxalate, Nitrate, and Lutein in Spinach

Released on J-STAGE: July 25, 2018 | Volume 24 Issue 3 Pages 421-425

Zheng Wang, Akira Ando, Atsuko Takeuchi, Hiroshi Ueda

Measurement of Water Absorption in Wheat Flour by Mixograph Test

Released on J-STAGE: December 28, 2016 | Volume 22 Issue 6 Pages 841-846

Reiko Okuda, Aya Tabara, Hideki Okusu, Masaharu Seguchi

Evaluation of the Antioxidant and Antimicrobial Activity of Rosemary Essential Oils as Gelatin Edible Film Component

Released on J-STAGE: April 26, 2019 | Volume 25 Issue 2 Pages 321-329

Walid Yeddes, Malgorzata Nowacka, Katarzyna Rybak, Islem Younes, Majdi Hammami, Moufida Saidani-Tounsi, Dorota Witrowa-Rajchert

Safety Evaluation of 6-(Methylsulfinyl) Hexyl Isothiocyanate (6-MSITC) and Wasabi Sulfinyl, a 6-MSITC-containing Supplement

Released on J-STAGE: December 27, 2020 | Volume 26 Issue 6 Pages 813-824

Isao Okunishi, Tomoe Yamada-Kato, Jiro Saito, De-Xing Hou

Nutritional Values and Functional Properties of House Cricket ( Acheta domesticus ) and Field Cricket ( Gryllus bimaculatus )

Released on J-STAGE: September 26, 2019 | Volume 25 Issue 4 Pages 597-605

Natteewan Udomsil, Sumeth Imsoonthornruksa, Chotika Gosalawit, Mariena Ketudat-Cairns

• Volume Vol 30 Vol 29 Vol 28 Vol 27 Vol 26 Vol 25 Vol 24 Vol 23 Vol 22 Vol 21 Vol 20 Vol 19 Vol 18 Vol 17 Vol 16 Vol 15 Vol 14 Vol 13 Vol 12 Vol 11 Vol 10 Vol 9 Vol 8 Vol 7 Vol 6 Vol 5
• Add to favorites
• Announcement alert
• New arrival alert
• X @jsfst_office

Food Science and Technology International, Tokyo

Register with J-STAGE for free!

Already have an account? Sign in here

• Search Menu
• Advance articles
• Editor's Choice
• Supplement Archive
• Article Collection Archive
• Author Guidelines
• Submission Site
• Open Access
• Call for Papers
• Why Publish?
• About Nutrition Reviews
• About International Life Sciences Institute
• Editorial Board
• Early Career Editorial Board
• Advertising and Corporate Services
• Journals Career Network
• Self-Archiving Policy
• Journals on Oxford Academic
• Books on Oxford Academic

Article Contents

Introduction, artificial intelligence, machine learning, and deep learning, previous studies of artificial intelligence in engineering, pharmacy, and medicine, history of analytical methods and artificial intelligence in food science and nutrition, machine learning in agriculture and the food industry, potential of artificial intelligence to assess foods and their bioactive components beyond basic nutritional properties, acknowledgements.

• < Previous

Artificial intelligence in food science and nutrition: a narrative review

Taiki Miyazawa and Yoichi Hiratsuka contributed equally to this work.

• Article contents
• Figures & tables
• Supplementary Data

Taiki Miyazawa, Yoichi Hiratsuka, Masako Toda, Nozomu Hatakeyama, Hitoshi Ozawa, Chizumi Abe, Ting-Yu Cheng, Yuji Matsushima, Yoshifumi Miyawaki, Kinya Ashida, Jun Iimura, Tomohiro Tsuda, Hiroto Bushita, Kazuichi Tomonobu, Satoshi Ohta, Hsuan Chung, Yusuke Omae, Takayuki Yamamoto, Makoto Morinaga, Hiroshi Ochi, Hajime Nakada, Kazuhiro Otsuka, Teruo Miyazawa, Artificial intelligence in food science and nutrition: a narrative review, Nutrition Reviews , Volume 80, Issue 12, December 2022, Pages 2288–2300, https://doi.org/10.1093/nutrit/nuac033

• Permissions Icon Permissions

In the late 2010s, artificial intelligence (AI) technologies became complementary to the research areas of food science and nutrition. This review aims to summarize these technological advances by systematically describing the following: the use of AI in other fields (eg, engineering, pharmacy, and medicine); the history of AI in relation to food science and nutrition; the AI technologies currently used in the agricultural and food industries; and some of the important applications of AI in areas such as immunity-boosting foods, dietary assessment, gut microbiome profile analysis, and toxicity prediction of food ingredients. These applications are likely to be in great demand in the near future. This review can provide a starting point for brainstorming and for generating new AI applications in food science and nutrition that have yet to be imagined.

A particular challenge for the fields of food science and nutrition is how to understand, organize, and use the various research outputs obtained through studies. One solution to this challenge is the use of artificial intelligence (AI). In recent years, significant developments in technologies using AI have been observed in the fields of engineering, pharmacy, and medicine. In the late 2010s, these technologies were complementary to the research areas of agriculture, food science, and nutrition. According to a report by the European Food Safety Authority, research needs for food safety regulations by 2030 are expected to include both intensified cooperation with society as a whole and increased information provided by society for risk assessment using AI (primarily focusing on machine learning [ML]) through real-time analysis of big data. 1 Foods are composed of various compounds, and a thorough understanding of how food functions within the body or which combinations of food components are optimal for human health requires an analysis of food compounds.

For many years, food and nutrition researchers have attempted to elucidate the mechanisms through which food compounds affect human nutrition, but time, resources, and cost have all presented major limitations. To overcome these challenges, it is expected that AI technology will be able to analyze data that would take many decades for researchers to analyze themselves, leading to great progress in this field. Against this background, the aim of this review is to systematically describe (1) the use of AI in the fields of engineering, pharmacy, and medicine, (2) the history of AI in relation to food science and nutrition, (3) AI technologies currently used in agriculture and the food industry, and (4) some of the important applications of AI that will likely be in great demand, especially in the future, in areas such as enhancement of immunity by foods, analysis of gut microbiome profiles, and toxicity prediction of food ingredients. It is hoped that this overview of AI use in the fields of food science and nutrition will provide a guide for developing new technologies.

Literature search strategy

The literature search was conducted using the PubMed, Google Scholar, and Web of Science databases. The following keywords were used: “quantitative structure-activity relationship,” “agriculture,” “analytical methods,” “artificial intelligence,” “big data,” “deep learning,” “dietary components,” “drug design,” “food function,” “food industry,” “food toxicity,” “generative adversarial network,” “gut microbiome,” “health,” “history,” “ image interpretation,” “immunity,” “machine learning,” “medical,” “molecular design,” “neural networks,” “nutritional science,” “pharmacological,” “prediction,” “reinforcement learning,” “representation learning,” “supervised learning,” “toxicity,” and “unsupervised learning.” For literature on the application of AI technology to foods, reports from 2020 or later were selected unless it was not possible to find a paper published within this time frame. The selected literature was first evaluated by all authors independently. Subsequently, after discussion among all authors, only those publications that were highly rated were selected. Publications were rated on the basis of the helpfulness of the literature for future applications of AI in the fields of food science and nutrition.

Artificial intelligence is not a single monolithic term but a relatively broad concept, defined by Baker and Smith (in the review by Zawacki-Richter et al 2 ) as “computers which perform cognitive tasks, usually associated with human minds, particularly learning and problem-solving.” According to them, the definition of AI does not refer to a single technology, since the term covers a variety of technologies such as ML, deep learning (DL), data mining, and neural networks. Machine learning is a subfield of AI and can be described as an AI algorithm that extracts patterns from raw data to make subjective decisions. Machine learning was defined by Popenici and Kerr 3 in 2017 as “as a subfield of artificial intelligence that includes software able to recognise patterns, make predictions, and apply newly discovered patterns to situations that were not included or covered by their initial design.” Deep learning is a subfield of ML. Ayturan et al 4 described 2 differences between ML and DL. The first is the difference in data size. Unlike ML, DL uses big data. Furthermore, DL focuses on end-to-end problem solving, whereas ML uses problem division and management techniques. The second difference is that ML is processed in a single layer, while DL can be processed in multiple successive layers simultaneously. There have been remarkable developments in AI, ML, and DL in recent years, which have already been applied to a number of research fields. The next section will present an overview of previous studies on the application of AI in fields other than food science and nutrition.

Engineering

Fifty years after the term artificial   intelligence was first used at the Dartmouth Conference in 1956, 5 the first breakthrough papers on DL were published by Hinton et al 6 and Hinton and Salakhutdinov 7 in 2006. Originally, DL was an ML method that used multilayered neural networks, but today it refers more abstractly to all AI techniques that learn complex concepts by hierarchizing simple concepts. 8–10

In 2012, a group from Google published a famous paper, known as the “Google Cat Paper,” on successful image recognition of a cat’s face using unsupervised DL. 11 In the same year, Hinton’s research group, Krizhevsky et al, 12 , 13 won the ImageNet Large Scale Visual Recognition Challenge, a general object recognition contest, using a deep convolutional networks that outperformed other methods that use DL. 12 , 13 This represented a significant achievement because it demonstrated that DL can automatically acquire the features of an image without needing humans to extract them. Moreover, only 3 years later, in 2015, Microsoft’s group 14 and Google’s group 15 successively achieved image recognition capabilities that surpassed the human error rate of 5.1%. This is partially because of improvements in computational capabilities, such as graphics processing units with a large number of cores and tensor processing units optimized for DL. Since then, research on image and speech recognition using DL has increased explosively, the performance of DL has improved drastically, and the application of AI in engineering fields has spread rapidly. 16 , 17 Examples of these practical applications include natural language processing (NLP) systems (eg, Google’s BERT), 18 language translation systems that use neural machine translation (NMT), 19 , 20 voice assistant systems (eg, Apple’s Siri), 21 recommender systems (eg, IBM’s Watson), 22 , 23 and AI demand forecasting systems (eg, NEC’s Supply and Demand Optimization Platform). 24

Additionally, although they are still in the research stage, molecular-level material design and development, 25 material property prediction, 26 and protein structure prediction 27 , 28 are important applications in the evolution of ML models. In particular, AlphaFold 2, a protein structure prediction AI program developed by DeepMind Technologies (a London-based subsidiary of Google), had a tremendous impact on the 14th Critical Assessment of protein Structure Prediction (CASP) in 2020, showing prediction accuracy comparable to experimental techniques such as x-ray crystallography. AlphaFold 2 also predicted several protein structures of SARS-CoV-2, one of which, the ORF3a protein structure, was in good agreement with the experimental results of cryo-electron microscopy that were published later. 27

In 2016, AlphaGo, a computer Go program developed by DeepMind, defeated world-class Go players using deep reinforcement learning, a combination of DL and reinforcement learning. 29 , 30 This has since evolved into AlphaZero, a single program that automatically learns to play various games other than Go (eg, chess and shogi). 31 , 32 At present, deep reinforcement learning applications are limited to the field of robotics, especially in autonomous driving, 33 , 34 and thus further applications are expected.

In 2014, Goodfellow et al 35 published a Generative Adversarial Network (GAN), a method for generating new pseudo-images based on features automatically extracted from prepared image data. Since then, GAN-related research has expanded GAN’s applications to speech processing and natural language processing. Examples include super-resolution, which converts rough images to high-quality images from natural language, and music generation. 36

Pharmacy and medicine

Recently, the use of AI technology that is capable of processing big data has attracted attention and presented new challenges in the fields of pharmacy and medicine. Artificial intelligence technologies in medicine are used for diagnostics and disease prevention (through early identification of disease patterns and signs). Meanwhile, AI technologies in the pharmaceutical field are used in clinical trials, drug design, formulations of pharmaceutical preparations, and electronic prescriptions. 37 , 38 Schneider and Clark 39 reviewed recent examples of automated de novo molecular design, discussed the concepts and computational approaches involved, and predicted of some of the possibilities and limitations of drug design using machine intelligence.

Machine learning has been explored for predicting successful treatment with antidepressant medication, 40 , 41 for characterizing depression, 42–44 and for predicting suicide. 42 , 45–47 Deep learning has been applied to drug design and the prediction of drug–target interactions, drug toxicity, the biological activity of drugs, and the pharmacological properties of drugs. 48 , 49 Generative tensorial reinforcement learning (GENTRL), a type of DL, has been developed for rapid and effective molecular design and has been used to generate candidate compounds of potent inhibitors of discoidin domain receptor 1. 49

As mentioned previously, AI technology has been used for diagnosis and prediction. 40 In radiology, there have been studies on the diagnosis of acute neurological events, 50 brain hemorrhage, 51 head trauma, 52 chest disease, 53 , 54 breast disease, 55 and wrist fracture 56 using AI technology and images of computed tomography, chest radiography, and so on. Artificial intelligence technology has also been applied to the fields of pathology, 57–61 dermatology, 62 , 63 ophthalmology, 64–71 gastroenterology, 72 , 73 and cardiology. 74 , 75 . In particular, the US Food and Drug Administration has approved several proprietary algorithms for image interpretation in the field of pathology, and many companies have been involved. 40 In 2018, the US Food and Drug Administration published a fast-track approval plan for AI medical algorithms. Moreover, a virtual medical coach model that can provide virtual health guidance has been proposed. 40

Artificial intelligence technology can be applied not only to diagnosis but also to data analysis. Machine learning has been used in the analytics of genomics and other omics biology data sets. Open-source algorithms have been developed for classification or analysis of whole-genome sequence pathogenic variants, 76–81 somatic cancer mutations, 82 gene–gene interactions, 83 RNA sequencing data, 84 DNA methylation states, 85 protein–protein interactions, 86 and microbiomes 87 as well as for prediction of protein structures 88 and improvement of single-cell RNA sequencing. 89 While these reports present a single-cell omics approach, multi-omics algorithms that integrate the data set have been developed. 90 Algorithmic prediction of CRISPR guide RNA activity 91 and off-target activities 92 has also aided in the use of genome editing. Another potential area in which AI technology might be used in medicine is robotics. Medical robots can be used during surgery and in the delivery of care for elderly patients. 93 The AI methods described above are gradually being considered for use in the fields of food science and nutrition. One reason that AI technology has developed in other areas while its application to food science and nutrition has been relatively slow is probably related to the fact that food is composed of various compounds, as mentioned in the Introduction. Moreover, although the physiological functions of food are considered closely related to metabolism, a 2019 editorial in Nature Metabolism noted that the application of AI in metabolism research is far less advanced than that in other biomedical and life science fields, such as neuroscience and genomics. 94 This was attributed to the lack of high-quality data sets, which are essential for the successful application of platforms such as ML algorithms, in metabolism research.

As technology has evolved, so too have analytical methods for various components of food ( Figure 1 , see Development of nutrition and food chemistry and Development of food analytical methods ). In the 1930s, various analytical methods were developed, including gas chromatography, thin-layer chromatography, and pH meters. 95 From the 1940s to the 1970s, analytical methods were developed that are still widely used today. Since 1980, analytical methods such as enzyme-linked immunosorbent assay, inductively coupled plasma-mass spectrometry, capillary electrophoresis mass spectrometry, and several soft ionization techniques (fast atom bombardment, electrospray ionization, and matrix-assisted laser desorption/ionization) have been developed and are still used and recognized to be highly accurate. 96 The development of these methods has enabled the elucidation of new food functions.

Overlapping timelines of the development of nutrition and food chemistry, food analytical methods, and artificial intelligence (AI). Since the late 2010s, these three fields have developed a powerful synergy. These are representative innovations, not intended to be comprehensive.

Against the background of these discoveries in food science and nutrition and the development of new analytical methods, there have been remarkable advances in AI in recent years ( Figure 1 , see Development of artificial intelligence ). Technologies and gadgets based on these techniques are becoming commonplace in human life. Chronologically, the first research on AI was published in 1943 by McCulloch and Pitts, 97 under the title “A Logical Calculus of the Ideas Immanent in Nervous Activity.” This study described a simplified model of a network of neurons and the execution of logical functions. Additionally, the approach in this report has implications for developing current computer-based neural network models. In 1957, Rosenblatt 98 developed the perceptron algorithm, which modeled vision and the brain as a functional neural network, enabling pattern recognition. In 1964, Bobrow 99 conducted a study entitled “Natural Language Input for a Computer Problem Solving System” and developed STUDENT, a computer program for natural language understanding. In 1969, Minsky and Papert 100 published their book Perceptrons: An Introduction to Computational Geometry , which details the definition of perceptron and describes pattern recognition of the perceptron, parity and connectedness, and exclusive disjunction (also called XOR). They also reported the limitations of simple neural networks. In 1989, LeCun et al 101 created a backpropagation algorithm in a multilayer neural network that could recognize handwritten postal codes. In the 1990s, recurrent neural networks and support vector machines were proposed, which made it possible to show temporal dynamic behavior for time series. 102 In 2006, Hinton 103 published Learning Multiple Layers of Representation , which reported an approach that led to the DL subset of ML. As mentioned in the section “Previous Studies of the Application of Artificial Intelligence in Engineering, Pharmacy, and Medicine,” technologies established using AI developments are becoming essential in the fields of engineering, pharmacy, and medicine.

In the latter half of 2010, discoveries in nutritional science, food chemistry, food analytical methods, and AI, which had been developed separately in the past, began to overlap. For example, AI has been adapted to nutritional epidemiology (eg, dietary pattern analysis and detection of foodborne illness using big data), 104 , 105 food toxicity assessment, 106 , 107 image diagnosis, 108 , 109 and personalized nutrition 110 , 111 ( Figure 1 , see Combination of food chemistry, analytical chemistry and artificial intelligence ). Ramyaa et al 112 used ML to conduct nutritional phenotyping, demonstrating another use of AI that has applications for personalized nutrition. Their ML model predicted the current weight (in kilograms) and body mass index of participants in the Women's Health Initiative Observational Study using nutritional epidemiological data such as intakes of carbohydrates, protein, fat, fiber, and sugar, along with physical activity variables. Their results suggest that ML for personalizing dietary intake levels may be useful at the population level. Nakamura et al 113 reported a framework that can provide personalized health improvement plans based on the predictions of ML models designed to compute processes for health improvement. Berry et al 114 developed a model to predict serum triglycerides and glucose tolerance based on primary data from a British cohort study (n = 1002 participants). Their results were validated independently in an American cohort (n = 100), in which their prediction model was found to predict serum triglycerides and glucose. Furthermore, the results showed that there is considerable individual variability in postprandial metabolic responses to the same diet. Guasch-Ferré 115 et al analyzed 385 walnut-derived metabolites in human plasma using liquid chromatography coupled with tandem mass spectrometry and attempted to clarify the relationship between these metabolites and the risk of type 2 diabetes and cardiovascular disease using ML (Pearson correlation coefficients were assessed between metabolite-weighted models and self-reported walnut consumption in each pair of training–validation data sets within the discovery population). The results showed that 19 metabolites associated with walnut consumption were associated with a reduced risk of type 2 diabetes and cardiovascular disease.

In 2020, the AI Diatrofi Subcommittee was established by the International Life Sciences Institute Japan (ILSI Japan). This subcommittee aims to establish a predictive model for the interaction between the intake of multimolecular foods and human health using AI technology. 116 Furthermore, this subcommittee is attempting to use AI technology to develop various novel foods that humans have not yet considered.

As noted in the Introduction, several applications of AI have recently begun to be used in agriculture, food science, nutrition, the food industry, and related economies. The most widely studied AI technique is ML, which can be broadly classified into 4 categories: reinforcement learning, representation learning, supervised learning, and unsupervised learning ( Figure 2 ). 117–132

Examples of machine learning (ML) classifications and   applications relevant to the agricultural and food industries.

Reinforcement learning is an algorithm in which AI achieves a goal in an uncertain and potentially complex environment to build an ML model for decision-making ( Figure 2 ). The method then aims to maximize the total reward by obtaining rewards and penalties for the actions taken. Dharmasena et al 117 reported that the climate and irrigation in greenhouses can be automatically controlled by robots in farm and factory environments that can monitor temperature, soil moisture, humidity, and pH conditions and to detect unhealthy plants using image processing.

Representation learning is an ML method in which raw input data is fed into the system and AI is then allowed to automatically discover the expressions needed to achieve a goal ( Figure 2 ). To automate the detection of diseases and defects in plants, Abdu et al 118 proposed a novel ML approach that improves the identification performance without requiring optimization algorithms to select features, which often leads to redundancy in results. To achieve this, they employed pathological segmentation and localized feature extraction to minimize feature redundancy and feature vector length. To predict gene expression, Braun and Lawrence-Dill 119 computationally translated the natural language description of plant phenotypes into a structured representation, which enabled the generation of large-scale data sets without manual curation, leading to the identification of phenotypic similarities and the prediction of gene function within and across species. Overweg et al 120 proposed a reinforcement learning environment using a process-based crop growth model, which is helpful for optimizing fertilizer management strategies to reduce the detrimental effect of nitrogen fertilizers on the environment.

Supervised learning is a technique that uses labeled data sets to train algorithms to predict unknown outcomes ( Figure 2 ). To extract food functions, Zhao et al 121 reported that the application of a soft voting algorithm to seven ML algorithms significantly improved the accuracy of predicting food molecules with potential anti-cancer properties from 82% to 87%. For recommender systems, Kim and Chung 122 reported that a knowledge-based hybrid decision model using a neural network for nutrition management enables individualized dietary recommendations to be made by considering a person’s lifestyle, health conditions, and preferences. This model was also able to overcome problems associated with the current model. For targeted marketing in the food industry, Ali et al 123 reported that a new methodology for analyzing consumer preferences and acceptance of food flavors through facial emotion recognition provides a good prediction of sensory evaluation. To visualize big data, Aravind and Sweetlin 124 developed a new method for classifying food into 5 levels of healthiness that can make nutritional information more easily accessible. Lim et al 125 reported that supervised learning combined with target lipidomics can rapidly authenticate adulterated admixtures from original white rice samples in a short time.

Unsupervised learning is a method for analyzing and clustering unlabeled data sets using ML algorithms ( Figure 2 , Unsupervised learning). Gjoreski et al 126 reported that dietary intake data collected using a food propensity questionnaire and dependent 24-hour recalls were analyzed with unsupervised learning, which resulted in the identification of 4 clusters. The large national sets of nutrition-related data available in many countries should make it easier to create sustainable policies for agriculture, human health, and the environment. Babajide et al 127 reported that ML algorithms successfully categorized the weight change of individuals in a cohort study. The algorithms identified patterns in the effect of the type of diet on weight change, justifying the finding that a low-fat hypoenergetic diet resulted in weight loss. Aulia et al 128 reported that a model to predict the macronutrient content of baby food was developed using a DL approach based on the spectral profile of the food obtained using near-infrared spectroscopy. The results showed the model to perform most accurately for the identification of carbohydrate, protein, and fat.

The use of ML in the fields of food science and nutrition is often aimed at improving productivity and efficiency in the food industry, yet several challenges remain. Currently, ML models aimed at developing reliable and low-cost methods for food product quality control are being investigated. 133 Several studies using ML to ensure the safety and health compliance of employees working in food factories are under way. As an example of this type of study, AI-equipped closed-circuit TV cameras used facial recognition and object recognition software to determine whether workers met the personal hygiene requirements set forth in food safety laws. 134 Furthermore, if a worker was observed to be noncompliant with the requirements of food safety laws, the image on the screen could be extracted and checked in real time. 134 The freshness of fruits and vegetables is important for evaluating the quality of food, yet manual sorting by humans, which relies heavily on human judgment and leads to errors, reduced sorting efficiency, and increased costs, remains the most widely used method of sorting. 129 To address this problem, sensors that check the freshness of fruits and vegetables using ML can be combined, allowing sorting to be partially automated. By introducing AI into sensor-based optical sorting solutions, ML can automate the expression patterns for food freshness and enable highly efficient food sorting. 135 Machine learning is also being tested in the food processing industry, for example, to reduce the cost of food waste disposal. An automatic cleaning system (cleaning-in-place) for food processing equipment was programmed to clean the equipment in a timed cycle. This process begins with prerinsing and is followed by washing with an alkaline solution, intermediate rinsing, washing with an acidic solution, and final rinsing. 136 The existing cleaning-in-place system is a major obstacle in practical applications because it is a time-consuming and inefficient process that wastes large amounts of power, water, and chemicals. 137 However, by incorporating ML into a cleaning-in-place system, it has become possible to measure food residues and microbial debris in food processing equipment using devices equipped with UV light fluorescence imaging and ultrasonic acoustic sensors to optimize the cleaning process. 130 , 131

The use of ML in food supply chain management (a management method for building an integrated logistics system among multiple companies to increase efficiency, rather than limiting the logistics system to within 1 company) is also being attempted. To investigate the current state of research on AI applications in the supply chain, Riahi et al 138 reviewed 136 research papers in the Scopus database published between 1996 and 2020. They found that the number of publications in this area has been increasing year after year, revealing that, among three ML algorithms (ie, supervised, unsupervised, and reinforcement learning), reinforcement learning has received the most attention, especially in supply chain management. Using ML to monitor and test food safety at each stage in the supply chain has the potential to improve supply chain operations, ensure compliance with industry specifications, and meet consumer expectations. 132 Specifically, image recognition technology using ML enables more efficient procurement of agricultural products. 139 As already mentioned, many practical applications of AI are close to being viable in the food and agriculture industries, and some are already in use. In addition, other important applications of AI will likely be in great demand in the future, such as those used for gut microbiome profile analysis, toxicity prediction of food ingredients, and the development of foods with immunity-boosting properties. These topics are introduced in the next section.

Assessing the function of foods and their bioactive components

Artificial intelligence in nutrition has received significant attention. Some dietary components impart various health benefits beyond basic nutrition by modifying gastrointestinal function, promoting changes in biochemical parameters, and subsequently interacting with the body systems. Such bioactive components are expected to enhance biological defense mechanisms and prevent certain diseases. However, it takes years and involves tremendous costs to identify bioactive components in foods, to validate their efficacy and safety via human interventional studies, and to elucidate their mechanism of action. Moreover, most current research in food science and nutrition has investigated the biochemical and physiological properties of only a single food-derived component, even though dietary components are ingested together with more than thousands of food components that affect subsequent physiological responses additively, synergistically, or antagonistically. Artificial intelligence technology has the potential to elucidate the complex physiological mechanisms involved in the multivariate interactions between dietary components, thereby advancing current knowledge in nutritional science ( Figure 3 ).

Scheme of artificial intelligence (AI), machine learning (ML),   and deep learning (DL) for the understanding of “Food.”

Artificial intelligence is also expected to be used to predict the effect of whole foods or bioactive food components on health, ideally according to each individual’s physiological state. For example, previous studies of constipation and diarrhea have suggested that dietary intake, physical activity, and stress are associated with stool consistency. However, the effects of these factors are still unresolved and have not been investigated in healthy adults. Against this background, Lemay et al 140 investigated which factors (diet, physiology, lifestyle, and stress) affected stool hardness in 364 healthy adults, using ML data and statistical and ML analysis. They reported that the results of ML analysis can explain why stool hardness scored by technicians varies with dietary intake and stress hormones, but not physical activity. Furthermore, using dietary, physiological, and lifestyle data, their group applied several ML models (ridge, lasso, elastic net, and random forest models) to predict the bone mineral content and bone mineral density of the whole body, the femoral neck, and the spine in healthy adults (n = 313). 141 The ML models showed superior performance compared with multivariate linear regression. In conclusion, the study demonstrated that modifiable factors such as diet have less variability in the data compared with nonmodifiable factors such as age, sex, and ethnicity. Furthermore, lower fecal pH was shown to be associated with higher bone mineral content and bone mineral density.

The COVID-19 pandemic has raised awareness about the immunomodulatory functions of dietary components. Patients with noncommunicable diseases, such as cardiovascular disease, diabetes, and cancer, are more susceptible to COVID-19 infection. 142 The onset of noncommunicable diseases is triggered by the adverse biological responses of the immune system. In addition, the immune system undergoes dramatic changes with increasing age. These changes progress continuously, leading to a state of immune senescence as well as increased production of proinflammatory cytokines. 143 Therefore, it is important to identify and screen dietary components with immunomodulatory functions that lead to better health and antiaging effects and to elucidate their mechanism of action. Using ML and in silico predictions, Rein et al 144 discovered immunomodulatory peptides from rice protein and tested them for bioefficacy.

Chin et al 145 used AI to map a food composition database, using lactose as an example of a food component. They developed nine ML models to easily check the National Coordinating Center’s Nutrition Data System for Research using ML and database matching methods. Of the ML models tested, the XGB-Regressor model showed the best performance. The use of the descriptors “nutrient + text” provided the best estimates of lactose. These results suggest that the estimation of nutrients found exclusively in the Nutrition Coordinating Center’s database, which is time consuming to check manually, can be estimated more quickly by ML.

Nutrition is an important factor that determines the immune response. Many dietary components exert immunological functions by modulating cellular events in antigen-presenting cells, such as dendritic cells and macrophages. 146 However, reports on AI applications in antigen-presenting cells are limited. The activation, differentiation, and function of antigen-presenting cells are coordinated using signal transduction and metabolic pathways, as observed in many other cells. Several studies have shown that the combination of ML and abundant multi-omics data (transcriptome, proteomics, and metabolomics) can be used to effectively predict the dynamics and function of cellular pathways in an automated fashion. 90 , 147 Such analysis could be a tool to explain the mechanism of immunomodulation through dietary components. Artificial intelligence models mirroring the dynamics of intracellular events, including activation of signal transduction and metabolic pathways linked to cellular events, could be useful as the first step to predict the function of dietary components. Interestingly, DL models trained on a curated data set of 19 299 unique bacterial glycans have been established and shown to predict several glycan functions, including immunogenicity. 148 Such models can also be used to assess the immunogenicity of dietary glycans.

Dietary assessment software based on artificial intelligence

One concern about the evaluation of dietary assessment is the reliance on self-reports, which are prone to bias. To address these issues, Fazzino et al 149 evaluated the utility of the SmartIntake application. Briefly, SmartIntake and online dietary recalls were compared by (1) quantifying the ability of SmartIntake to monitor drinking behavior, (2) evaluating the usability of SmartIntake by the Computer System Usability Questionnaire, (3) conducting interviews, and (4) assessing the usage of SmartIntake, the preference for using SmartIntake vs recalls, and the consumption of alcohol. SmartIntake was demonstrated to be adaptable for measuring alcohol consumption in young adults, as it was found to capture 87% of all drinking occasions among college students (n = 15) who endorsed a pattern of heavy drinking.

In the study of AI in dietary assessment, Mezgec and Seljak 150 described 2 approaches to using DL models focused on meal evaluation with image-based approaches: (1) using already existing deep neural network definitions and (2) defining a new DL architectural approach. They reported 2 food image segmentation methods that rely on full convolutional networks and deep residual networks as a new DL architectural approach. Lu et al 151 developed a deep neural network–based meal evaluation system (goFOOD TM ) that can predict the calorie and macronutrient contents of food by using a 3D reconstruction algorithm. They reported that goFOOD TM performed better than an experienced nutritionist in estimating the composition of meals from a database of nonstandardized meals and performed as well as a nutritionist when tested on a database of fast foods. Boushey et al 152 studied the Mobile Food Record, an image-based dietary assessment method for mobile devices. This algorithm allows users to capture food intake and estimate energy and nutrient intake using a camera with a mobile device. Furthermore, they studied 45 healthy adult males and females between the ages of 21 and 65 and concluded that the Mobile Food Record is a user-friendly method whose accuracy is comparable to that of traditional dietary records and other image-based algorithms.

Assessing the links between the gut microbiome and individual-specific responses

Dietary components also modulate physiological responses by influencing the gut microbiome. The gut microbiome digests dietary components and produces various metabolites, such as short-chain fatty acids, polyamines, polyphenols, and vitamins. 153 These metabolites affect physiological responses by reprogramming the genome and altering epigenetic processes and/or by modulating gene expression and metabolic responses subsequently through receptor binding or transport to immune system cells, such as leukocytes and other body systems. 154 , 155 Machine learning has been applied to predict age, drug efficacy, and disease onset based on the relative abundance of microbes within the sampled microbiomes. 154 , 156

Machine learning models for analyzing changes in the gut microbiome and metabolites after ingestion of a certain type of dietary component (eg, prebiotics) have also been reported. 157 , 158 Zeevi et al 159 continuously monitored blood glucose levels for 1 week in a cohort of 800 individuals and measured the responses to 46 898 meals. They also designed an ML algorithm that integrated blood parameters, eating habits, anthropometric measurements, physical activity, and gut microbiota measured in their case study to accurately predict an individual's postprandial glycemic response to actual meals. Other reviews describe the use of AI to predict the health effects of food. 160–162 Shinn et al 163 used ML models (random forest model and out-of-bag estimation) to identify biomarkers and predict the effect of dietary intake on the human gut microbiota from the composition of fecal bacteria. They were able to predict the dietary intake of certain model foods (almonds, avocados, broccoli, walnuts, whole-grain barley, and whole-grain oats) with 70% to 85% accuracy. Such reports of the accuracy of ML models, even when input with complex fecal bacterial data, highlight the great potential of AI in the future. Such models could be part of an AI prediction model for the correlation between the gut microbiome and the effect of dietary components on health. Matusheski et al 164 discussed the potential of DL architecture to measure amounts of dietary components by food image recognition and thereby predict the impact of diet on the microbiome and subsequent physiological responses.

Assessing the toxicity of foods and their components

The use of AI in food toxicology has seen considerable advances. Traditional safety assessment of foods includes animal testing, but there are increasing demands to replace, reduce, and refine animal testing for this purpose. Several studies to predict toxicities have developed quantitative structure–activity relationship models using deep neural networks based on a database of chemical structural parameters and toxicities. 165–168 Quantitative structure–activity relationship models are also being used to predict the toxicity of nanoscale-processed food, called smart food. 169 Interestingly, Kozawa et al 170 designed a unique humanized mouse database, individualized (hMDB-i) using ML algorithms, drug-induced 24-organ mouse transcriptome patterns, and human clinical outcome data sets, aiming to predict drug adverse events and therapeutic indications. The efficacy and toxicity of compounds can be affected not only by their chemical and physical properties but also by individual genes involved in absorption, distribution, metabolism, and excretion. Ideally, a future AI model of food toxicology should reflect the individual absorption, distribution, metabolism, and excretion of dietary components.

Currently, various technologies that use AI are being developed in the fields of food science and nutrition. This review shows the potential of AI to make significant contributions to the fields of food science and nutrition. As discussed in the sections “Machine Learning in Agriculture and the Food Industry” and “Potential of Artificial Intelligence to Assess Foods and Their Bioactive Components Beyond Basic Nutritional Properties,” it can be inferred that AI will not only accelerate the efficiency of food production and processing but will also be useful in dietary assessment, development of food components with immunomodulatory properties, measurement of the effects of dietary components on the intestinal microbiota, and safety assessment of foods. Research has also begun to focus on the development of related technologies, as seen by the formation of the AI Diatrofi Subcommittee of the International Life Science Institute Japan, which aims to establish predictive models for the interaction between the simultaneous intake of multimolecular food products and human health. However, such studies have only recently commenced. Hence, continued research is required to obtain reliable information in the future. The fields of food science, nutrition, food analytical methods, and AI will continue to complement each other. This review is intended to be of interest to people in all of these fields in the hope it will serve as a catalyst for developing new technologies with multidisciplinary applications.

Author contributions. Ta.M., Y.H., M.T., N.H., H.O., and Te.M. conceived the review, organized the sections, and drafted the manuscript. Ta.M., M.T., N.H., and H.O. wrote the individual sections of the manuscript. C.A., T-Y.C., Y.Ma., Y.Mi., K.A., J.I., T.T., H.B., K.T., S.O., H.C., Y.O., T.Y., M.M., H.O., H.N., and K.O. critically reviewed the manuscript. All authors reviewed the final manuscript and agreed to submit it to this journal.

Funding. This paper was supported by ILSI Japan. The authors acknowledge Takuji Yasukawa, former president of ILSI Japan, and Hideyo Nakamura, former secretary general of ILSI Japan, for their support in establishing the AI Diatrofi Subcommittee in ILSI Japan. This study was supported by the Tohoku University Fund, Tohoku University, Sendai, Japan.

Declaration of interest. The authors have no relevant interests to declare.

Bronzwaer S , Kass G , Robinson T , et al. ; European Food Safety Authority (EFSA) . Food safety regulatory research needs 2030 . EFSA J.   2019 ; 17 : e170622 .

Google Scholar

Zawacki-Richter O , Marín VI , Bond M , et al.    Systematic review of research on artificial intelligence applications in higher education—where are the educators?   Int J Educ Technol High Educ.   2019 ; 16 : 19 .

Popenici SAD , Kerr S.   Exploring the impact of artificial intelligence on teaching and learning in higher education . Res Pract Technol Enhanc Learn.   2017 ; 12 : 22 .

Ayturan YA , Ayturan ZC , Altun OH.   Air pollution modelling with deep learning: a review . Int J Environ Pollut Environ Model . 2018 ; 1 : 58 – 62 .

McCarthy J , Minsky ML , Rochester N , et al.    A proposal for the Dartmouth Summer Research Project on Artificial Intelligence, August 31, 1955 . AI Magazine . 2006 ; 27 : 12 – 14 . doi: 10.1609/aimag.v27i4.1904

Hinton GE , Osindero S , Teh YW.   A fast learning algorithm for deep belief nets . Neural Comput.   2006 ; 18 : 1527 – 1554 .

Hinton GE , Salakhutdinov RR.   Reducing the dimensionality of data with neural networks . Science.   2006 ; 313 : 504 – 507 .

Bengio Y , Courville A , Vincent P.   Representation learning: a review and new perspectives . IEEE Trans Pattern Anal Mach Intell.   2013 ; 35 : 1798 – 1828 .

LeCun Y , Bengio Y , Hinton G.   Deep learning . Nature . 2015 ; 521 : 436 – 444 .

Goodfellow I , Bengio Y , Courville A.   Deep Learning . MIT Press; 2016 .

Le QV , Ranzato MA , Monga R , et al.  Building high-level features using large scale unsupervised learning. [abstract] In: Proceedings of the 29th International Conference on Machine Learning (ICML-12) . Edinburgh, UK: International Machine Learning Society; 2012 : 507 – 514 . https://dl.acm.org/doi/proceedings/10.5555/3042573

Krizhevsky A , Sutskever I , Hinton GE , ImageNet classification with deep convolutional neural networks [abstract]. In: Proceedings of the 25th International Conference on Neural Information Processing Systems . Volume 1. NeurIPS Foundation: Lake Tahoe, NV; 2012 : 1097 – 1105 . https://dl.acm.org/doi/proceedings/10.5555/2999134

Krizhevsky A , Sutskever I , Hinton GE.   ImageNet classification with deep convolutional neural networks . Commun ACM.   2017 ; 60 : 84 – 90 .

He K , Zhang X , Ren S , et al.  Delving deep into rectifiers: surpassing human-level performance on ImageNet classification [abstract]. In: Proceedings of the 2015 IEEE International Conference on Computer Vision (ICCV) . Institute of Electrical and Electronics Engineers (IEEE): Santiago, Chile; 2015 : 1026 – 1034 .

Ioffe S , Szegedy C.   Batch normalization: accelerating deep network training by reducing internal covariate shift [abstract] . In: Proceedings of the 32nd International Conference on Machine Learning (ICML-15) . International Machine Learning Society: Lille, France: 2015 ; 37 : 448 – 456 . https://dl.acm.org/doi/proceedings/10.5555/3045118

Injadat M , Moubayed A , Nassif AB , et al.    Machine learning towards intelligent systems: applications, challenges, and opportunities . Artif Intell Rev.   2021 ; 54 : 3299 – 3348 .

Sarker IH.   Machine learning: algorithms, real-world applications and research directions . SN Comput Sci.   2021 ; 2 : 160 .

Devlin J , Chang MW , Lee K.   BERT: pre-training of deep bidirectional transformers for language understanding [abstract] . In: Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies . Volume 1. Minneapolis, MN: 2019 ; 1 : 4171 – 4186 . https://aclanthology.org/volumes/N19-1/

Wu Y , Schuster M , Chen Z , et al.  Google's neural machine translation system: Bridging the gap between human and machine translation. arXiv . 2016 ;1609.08144. [10.48550/arXiv.1609.08144]

Johnson M , Schuster M , Le QV , et al.    Google's multilingual neural machine translation system: enabling zero-shot translation . Trans Assoc Comp Linguist.   2017 ; 5 : 339 – 351 .

Marchi E , Shum S , Hwang K , et al.  Generalised discriminative transform via curriculum learning for speaker recognition [abstract]. In: Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) . Calgary, Alberta; 2018 : 5324 – 5328 .

Ferrucci D , Brown E , Chu-Carroll J , et al.    Building Watson: an overview of the DeepQA project . AI Mag.   2010 ; 31 : 59 – 79 .

Ahmed MN , Toor AS , O'Neil K , et al.    Cognitive computing and the future of health care cognitive computing and the future of healthcare: the cognitive power of IBM Watson has the potential to transform global personalized medicine . IEEE Pulse.   2017 ; 8 : 4 – 9 .

Morita R , Yasuda K.   The supply and demand optimization platform: a co-creation project in value chains that reduces food losses . NEC Tech J . 2021 ; 15 : 18 – 22 .

Vasudevan R , Pilania G , Balachandran PV.   Machine learning for materials design and discovery . J Appl Phys . 2021 ; 129 : 070401 .

Chibani S , Coudert FX.   Machine learning approaches for the prediction of materials properties . APL Materials . 2020 ; 8 : 080701 .

Callaway E.   ‘It will change everything’: DeepMind’s AI makes gigantic leap in solving protein structures . Nature . 2020 ; 588 : 203 – 204 .

Service RF.   ‘The game has changed.’ AI triumphs at protein folding . Science.   2020 ; 370 : 1144 – 1145 .

Silver D , Huang A , Maddison CJ , et al.    Mastering the game of Go with deep neural networks and tree search . Nature . 2016 ; 529 : 484 – 489 .

Silver D , Schrittwieser J , Simonyan K , et al.    Mastering the game of Go without human knowledge . Nature . 2017 ; 550 : 354 – 359 .

Silver D , Hubert T , Schrittwieser J , et al.    A general reinforcement learning algorithm that masters chess, shogi, and Go through self-play . Science.   2018 ; 362 : 1140 – 1144 .

Schrittwieser J , Antonoglou I , Hubert T , et al.    Mastering Atari, Go, chess and shogi by planning with a learned model . Nature . 2020 ; 588 : 604 – 609 .

Grigorescu S , Trasnea B , Cocias T , et al.    A survey of deep learning techniques for autonomous driving . J Field Robotics.   2020 ; 37 : 362 – 386 .

Osiński B , Jakubowski A , Zięcina P , et al.  Simulation-based reinforcement learning for real-world autonomous driving [abstract]. In: Proceedings of the 2020 IEEE International Conference on Robotics and Automation (ICRA) . (Virtual); 2020 : 6411 – 6418 .

Goodfellow I , Pouget-Abadie J , Mirza M , et al.    Generative adversarial nets [abstract] . In: Proceedings of the 27th International Conference on Neural Information Processing Systems . Volume 2. NeurIPS Foundation: Montreal, Quebec: 2014 ; 2 : 2672 – 2680 . https://dl.acm.org/doi/proceedings/10.5555/2969033

Gui J , Sun Z , Wen Y , et al.  A review on generative adversarial networks: algorithms, theory, and applications. arXiv . 2020 . 2001.06937. [10.48550/arXiv.2001.06937]

Kostic EJ , Pavlović DA , Živković MD.   Applications of artificial intelligence in medicine and pharmacy: ethical aspects . Acta Medica Medianae . 2019 ; 58 : 128 – 136 .

Flynn A.   Using artificial intelligence in health-system pharmacy practice: finding new patterns that matter . Am J Health Syst Pharm.   2019 ; 76 : 622 – 627 .

Schneider G , Clark DE.   Automated de novo drug design: are we nearly there yet?   Angew Chem Int Ed Engl.   2019 ; 58 : 10792 – 10803 .

Topol EJ.   High-performance medicine: the convergence of human and artificial intelligence . Nat Med.   2019 ; 25 : 44 – 56 .

Chekroud AM , Zotti RJ , Shehzad Z , et al.    Cross-trial prediction of treatment outcome in depression: a machine learning approach . Lancet Psychiat . 2016 ; 3 : 243 – 250 .

Schnyer DM , Clasen PC , Gonzalez C , et al.    Evaluating the diagnostic utility of applying a machine learning algorithm to diffusion tensor MRI measures in individuals with major depressive disorder . Psychiatry Res Neuroimaging.   2017 ; 264 : 1 – 9 .

Reece AG , Danforth CM.   Instagram photos reveal predictive markers of depression . EPJ Data Sci.   2017 ; 6 : 15 .

Wager TD , Woo CW.   Imaging biomarkers and biotypes for depression . Nat Med.   2017 ; 23 : 16 – 17 .

Walsh CG , Ribeiro JD , Franklin JC.   Predicting risk of suicide attempts over time through machine learning . Clin Psychol Sci . 2017 ; 5 : 457 – 469 .

Franklin JC , Ribeiro JD , Fox KR , et al.    Risk factors for suicidal thoughts and behaviors: a meta-analysis of 50 years of research . Psychol Bull.   2017 ; 143 : 187 – 232 .

Just MA , Pan L , Cherkassky VL , et al.    Machine learning of neural representations of suicide and emotion concepts identifies suicidal youth . Nat Hum Behav.   2017 ; 1 : 911 – 919 .

Jing Y , Bian Y , Hu Z , et al.    Deep learning for drug design: an artificial intelligence paradigm for drug discovery in the big data era . AAPS J.   2018 ; 20 : 58 .

Zhavoronkov A , Ivanenkov YA , Aliper A , et al.    Deep learning enables rapid identification of potent DDR1 kinase inhibitors . Nat Biotechnol.   2019 ; 37 : 1038 – 1040 .

Titano JJ , Badgeley M , Schefflein J , et al.    Automated deep-neural-network surveillance of cranial images for acute neurologic events . Nat Med.   2018 ; 24 : 1337 – 1341 .

Arbabshirani MR , Fornwalt BK , Mongelluzzo GJ , et al.    Advanced machine learning in action: identification of intracranial hemorrhage on computed tomography scans of the head with clinical workflow integration . NPJ Digit Med.   2018 ; 1 : 9 .

Chilamkurthy S , Ghosh R , Tanamala S , et al.    Deep learning algorithms for detection of critical findings in head CT scans: a retrospective study . Lancet.   2018 ; 392 : 2388 – 2396 .

Nam JG , Park S , Hwang EJ , et al.    Development and validation of deep learning-based automatic detection algorithm for malignant pulmonary nodules on chest radiographs . Radiology . 2019 ; 290 : 218 – 228 . doi: 10.1148/radiol.2018180237

Singh R , Kalra MK , Nitiwarangkul C , et al.    Deep learning in chest radiography: detection of findings and presence of change . PLoS One.   2018 ; 13 : e0204155 .

Lehman CD , Yala A , Schuster T , et al.    Mammographic breast density assessment using deep learning: clinical implementation . Radiology . 2019 ; 290 : 52 – 58 .

Lindsey R , Daluiski A , Chopra S , et al.    Deep neural network improves fracture detection by clinicians . Proc Natl Acad Sci U S A.   2018 ; 115 : 11591 – 11596 .

Bejnordi BE , Veta M , Diest PJV , et al. ; the CAMELYON16 Consortium . Diagnostic assessment of deep learning algorithms for detection of lymph node metastases in women with breast cancer . JAMA . 2017 ; 318 : 2199 – 2210 .

Coudray N , Ocampo PS , Sakellaropoulos T , et al.    Classification and mutation prediction from non-small cell lung cancer histopathology images using deep learning . Nat Med.   2018 ; 24 : 1559 – 1567 .

Capper D , Jones DTW , Sill M , et al.    DNA methylation-based classification of central nervous system tumours . Nature . 2018 ; 555 : 469 – 474 .

Steiner DF , MacDonald R , Liu Y , et al.    Impact of deep learning assistance on the histopathologic review of lymph nodes for metastatic breast cancer . Am J Surg Pathol.   2018 ; 42 : 1636 – 1646 .

Liu Y , Kohlberger T , Norouzi M , et al.    Artificial intelligence-based breast cancer nodal metastasis detection: insights into the black box for pathologists . Arch Pathol Lab Med.   2019 ; 143 : 859 – 868 .

Haenssle HA , Fink C , Schneiderbauer R , et al. ; Reader Study Level-I and Level-II Groups . Man against machine: diagnostic performance of a deep learning convolutional neural network for dermoscopic melanoma recognition in comparison to 58 dermatologists . Ann Oncol.   2018 ; 29 : 1836 – 1842 .

Han SS , Kim MS , Lim W , et al.    Classification of the clinical images for benign and malignant cutaneous tumors using a deep learning algorithm . J Invest Dermatol.   2018 ; 138 : 1529 – 1538 .

Gulshan V , Peng L , Coram M , et al.    Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs . JAMA . 2016 ; 316 : 2402 – 2410 .

Abràmoff MD , Lavin PT , Birch M , et al.    Pivotal trial of an autonomous AI-based diagnostic system for detection of diabetic retinopathy in primary care offices . NPJ Digit Med.   2018 ; 1 : 39 .

Kanagasingam Y , Xiao D , Vignarajan J , et al.    Evaluation of artificial intelligence-based grading of diabetic retinopathy in primary care . JAMA Netw Open.   2018 ; 1 : e182665 .

Long E , Lin H , Liu Z , et al.    An artificial intelligence platform for the multihospital collaborative management of congenital cataracts . Nat Biomed Eng.   2017 ; 1 : 1 – 8 .

Fauw JD , Ledsam JR , Romera-Paredes B , et al.    Clinically applicable deep learning for diagnosis and referral in retinal disease . Nat Med.   2018 ; 24 : 1342 – 1350 .

Burlina PM , Joshi N , Pekala M , et al.    Automated grading of age-related macular degeneration from color fundus images using deep convolutional neural networks . JAMA Ophthalmol.   2017 ; 135 : 1170 – 1176 .

Brown JM , Campbell JP , Beers A , et al. ; Imaging and Informatics in Retinopathy of Prematurity (i-ROP) Research Consortium . Automated diagnosis of plus disease in retinopathy of prematurity using deep convolutional neural networks . JAMA Ophthalmol.   2018 ; 136 : 803 – 810 .

Kermany DS , Goldbaum M , Cai W , et al.    Identifying medical diagnoses and treatable diseases by image-based deep learning . Cell . 2018 ; 172 : 1122 – 1131.e9 .

Mori Y , Kudo S , Misawa M , et al.    Real-time use of artificial intelligence in identification of diminutive polyps during colonoscopy: a prospective study . Ann Intern Med.   2018 ; 169 : 357 – 366 .

Wang P , Xiao X , Brown JRG , et al.    Development and validation of a deep-learning algorithm for the detection of polyps during colonoscopy . Nat Biomed Eng.   2018 ; 2 : 741 – 748 .

Madani A , Arnaout R , Mofrad M , et al.    Fast and accurate view classification of echocardiograms using deep learning . NPJ Digit Med . 2018 ; 1 : 6 . [10.1038/s41746-017-0013-1]

Zhang J , Gajjala S , Agrawal P , et al.    Fully automated echocardiogram interpretation in clinical practice: feasibility and diagnostic accuracy . Circulation . 2018 ; 138 : 1623 – 1635 .

Sundaram L , Gao H , Padigepati SR , et al.    Predicting the clinical impact of human mutation with deep neural networks . Nat Genet.   2018 ; 50 : 1161 – 1170 .

Zhou J , Theesfeld CL , Yao K , et al.    Deep learning sequence-based ab initio prediction of variant effects on expression and disease risk . Nat Genet.   2018 ; 50 : 1171 – 1179 .

Zhou J , Troyanskaya O.   Predicting effects of noncoding variants with deep learning–based sequence model . Nat Methods.   2015 ; 12 : 931 – 934 .

Luo R , Sedlazeck FJ , Lam TW , et al.    A multi-task convolutional deep neural network for variant calling in single molecule sequencing . Nat Commun.   2019 ; 10 : 998 .

Poplin R , Chang PC , Alexander D , et al.    A universal SNP and small-indel variant caller using deep neural networks . Nat Biotechnol.   2018 ; 36 : 983 – 987 .

Riesselman AJ , Ingraham JB , Marks DS.   Deep generative models of genetic variation capture the effects of mutations . Nat Methods.   2018 ; 15 : 816 – 822 .

Wood DE , White JR , Georgiadis A , et al.    A machine learning approach for somatic mutation discovery . Sci Transl Med . 2018 ; 10 : eaar7939 . [10.1126/scitranslmed.aar7939]

Behravan H , Hartikainen JM , Tengström M , et al.    Machine learning identifies interacting genetic variants contributing to breast cancer risk: a case study in Finnish cases and controls . Sci Rep.   2018 ; 8 : 13149 .

Lin C , Jain S , Kim H , et al.    Using neural networks for reducing the dimensions of single-cell RNA-Seq data . Nucleic Acids Res.   2017 ; 45 : e156 .

Angermueller C , Lee HJ , Reik W , et al.    DeepCpG: accurate prediction of single-cell DNA methylation states using deep learning . Genome Biol.   2017 ; 18 : 67 .

AlQuraishi M.   End-to-end differentiable learning of protein structure . Cell Syst.   2019 ; 8 : 292 – 301.e3 .

Espinoza JL.   Machine learning for tackling microbiota data and infection complications in immunocompromised patients with cancer . J Intern Med.   2018 ; 284 : 189 – 192 .

Senior AW , Evans R , Jumper J , et al.    Improved protein structure prediction using potentials from deep learning . Nature . 2020 ; 577 : 706 – 710 .

Van Dijk D , Sharma R , Nainys J , et al.    Recovering gene interactions from single-cell data using data diffusion . Cell . 2018 ; 174 : 716 – 729.e727 .

Camacho DM , Collins KM , Powers RK , et al.    Next-generation machine learning for biological networks . Cell . 2018 ; 173 : 1581 – 1592 .

Kim HK , Min S , Song M , et al.    Deep learning improves prediction of CRISPR–Cpf1 guide RNA activity . Nat Biotechnol.   2018 ; 36 : 239 – 241 .

Listgarten J , Weinstein M , Kleinstiver BP , et al.    Prediction of off-target activities for the end-to-end design of CRISPR guide RNAs . Nat Biomed Eng.   2018 ; 2 : 38 – 47 .

Hamet P , Tremblay J.   Artificial intelligence in medicine . Metabolism . 2017 ; 69 : S36 – S40 .

Why the metabolism field risks missing out on the AI revolution . Nat Metab . 2019 ; 1 : 929 – 930 . [10.1038/s42255-019-0133-9]

Gallo M , Ferranti P.   The evolution of analytical chemistry methods in foodomics . J Chromatogr A.   2016 ; 1428 : 3 – 15 .

Olivares JA , Nguyen NT , Yonker CR , et al.    On-line mass spectrometric detection for capillary zone electrophoresis . Anal Chem.   1987 ; 59 : 1230 – 1232 .

McCulloch WS , Pitts W.   A logical calculus of the ideas immanent in nervous activity . Bull Math Biophys.   1943 ; 5 : 115 – 133 .

Rosenblatt F.   The perceptron: a probabilistic model for information storage and organization in the brain . Psychol Rev.   1958 ; 65 : 386 – 408 .

Bobrow DG. Natural language input for a computer problem solving system. Technical report. Massachusetts Institute of Technology; 1968 ; 133 – 215 . https://dl.acm.org/doi/book/10.5555/889266

Minsky M , Papert S.   Perceptrons: an Introduction to Computational Geometry . MIT Press; 1969 . 10.7551/mitpress/11301.001.0001

LeCun Y , Boser B , Denker JS , et al.    Backpropagation applied to handwritten zip code recognition . Neural Comput . 1989 ; 1 : 541 – 551 .

Li D , Du Y.   Artificial Intelligence with Uncertainty . 1st ed. Chapman and Hall/CRC; 2007 .

Hinton GE.   Learning multiple layers of representation . Trends Cogn Sci.   2007 ; 11 : 428 – 434 .

Zhang F , Tapera TM , Gou J.   Application of a new dietary pattern analysis method in nutritional epidemiology . BMC Med Res Methodol.   2018 ; 18 : 119 .

Sadilek A , Caty S , DiPrete L , et al.    Machine-learned epidemiology: real-time detection of foodborne illness at scale . NPJ Digit Med.   2018 ; 1 : 36 .

Li L , Rong S , Wang R , et al.    Recent advances in artificial intelligence and machine learning for nonlinear relationship analysis and process control in drinking water treatment: a review . Chem Eng J . 2021 ; 405 : 126673 .

Issac A , Srivastava A , Srivastava A , et al.    An automated computer vision based preliminary study for the identification of a heavy metal (Hg) exposed fish- channa punctatus . Comput Biol Med.   2019 ; 111 : 103326 .

Takahashi K , Doman K , Kawanishi Y , et al.  Estimation of the attractiveness of food photography focusing on main ingredients [abstract]. In: Proceedings of the 9th Workshop on Multimedia for Cooking and Eating Activities in conjunction with the 2017 International Joint Conference on Artificial Intelligence . Melbourne, Australia; 2017 : 1 – 6 . [10.1145/3106668.3106670]

Zhou L , Zhang C , Liu F , et al.    Application of deep learning in food: a review . Compr Rev Food Sci Food Saf.   2019 ; 18 : 1793 – 1811 .

Sempionatto JR , Montiel VRV , Vargas E , et al.    Wearable and mobile sensors for personalized nutrition . ACS Sens.   2021 ; 6 : 1745 – 1760 .

Smetana S , Aganovic K , Heinz V.   Food supply chains as cyber-physical systems: a path for more sustainable personalized nutrition . Food Eng Rev.   2021 ; 13 : 92 – 103 .

Ramyaa R , Hosseini O , Krishnan GP , et al.    Phenotyping women based on dietary macronutrients, physical activity, and body weight using machine learning tools . Nutrients   2019 ; 11 : 1681 .

Nakamura K , Kojima R , Uchino E , et al.    Health improvement framework for actionable treatment planning using a surrogate Bayesian model . Nat Commun.   2021 ; 12 : 3088 .

Berry SE , Valdes AM , Drew DA , et al.    Human postprandial responses to food and potential for precision nutrition . Nat Med.   2020 ; 26 : 964 – 973 .

Guasch-Ferré M , Hernández-Alonso P , Drouin-Chartier J-P , et al.    Walnut consumption, plasma metabolomics, and risk of type 2 diabetes and cardiovascular disease . J Nutr . 2021 ; 151 : 303 – 311 .

International Life Sciences Institute Japan (ILSI Japan) . Notice of establishment of new joint research course of ILSI Japan and recruitment of participation. ILSI Japan website. Published August 2020. Accessed June 13, 2021 . http://www.ilsijapan.org/ILSIJapan/COM/AI/AI2020.php

Dharmasena T , Silva R , Abhayasingha N , et al.  Autonomous cloud robotic system for smart agriculture. In: Proceedings of the 2019 Moratuwa Engineering Research Conference (MERCon) . Moratuwa, Sri Lanka; 2019 : 388 – 393 .

Abdu AM , Mokji MM , Sheikh UU.   Automatic vegetable disease identification approach using individual lesion features . Comput Electron Agric . 2020 ; 176 : 105660 .

Braun IR , Lawrence-Dill CJ.   Automated methods enable direct computation on phenotypic descriptions for novel candidate gene prediction . Front Plant Sci.   2019 ; 10 : 1629 .

Overweg H , Berghuijs HNC , Athanasiadis IN , et al.  CropGym: A reinforcement learning environment for crop management. arXiv . 2021 :2104.04326.

Zhao S , Mao X , Lin X , et al.    Machine learning prediction for 50 anti-cancer food molecules from 968 anti-cancer drugs . Int J Intell Sci.   2020 ; 10 : 1 – 8 .

Kim J-C , Chung K.   Knowledge-based hybrid decision model using neural network for nutrition management . Inf Technol Manag.   2020 ; 21 : 29 – 39 .

Ali T , Ahmad I , Rehman AU , et al.  Understanding customer experiences through social media analysis of three giants of soft drink industry [abstract]. In : Proceedings of the 5th International Conference on Behavioral, Economic, and Socio-Cultural Computing (BESC) . Kaohsiung, Taiwan; 2018 .

Aravind J , Sweetlin JD , Nutrient facts analysis using supervised learning approaches [abstract]. In: Proceedings of the 2017 3rd Conference on Information and Communication Technology (CICT) . Gwalior, India; 2017 .

Lim DK , Long NP , Mo C , et al.    Combination of mass spectrometry-based targeted lipidomics and supervised machine learning algorithms in detecting adulterated admixtures of white rice . Food Res Int.   2017 ; 100 : 814 – 821 .

Gjoreski M , Kochev S , Reščič N , et al.  Exploring dietary intake data collected by FPQ using unsupervised learning [abstract]. In: Proceedings of the 2019 IEEE International Conference on Big Data (Big Data) . Los Angeles, CA; 2019 .

Babajide O , Tawfik H , Palczewska A , et al.  Application of unsupervised learning in weight-loss categorisation for weight management programs [abstract]. In: Proceedings of the 2019 10th International Conference on Dependable Systems, Services and Technologies (DESSERT) . Leeds, UK; 2019 .

Aulia MN , Khodra ML , Koesoema AP.   Predicting macronutrient of baby food using near-infrared spectroscopy and deep learning approach . IOP Conf Ser Mater Sci Eng.   2020 ; 803 : 012019 .

de Luna RG , Dadios EP , Bandala AA. Tomato fruit image dataset for deep transfer learning-based defect detection [abstract]. In: Proceedings of the 2019 IEEE International Conference on Cybernetics and Intelligent Systems (CIS) and the IEEE Conference on Robotics, Automation and Mechatronics (RAM) . Bangkok, Thailand; 2020 .

Escrig JE , Simeone A , Woolley E , et al.    Ultrasonic measurements and machine learning for monitoring the removal of surface fouling during clean-in-place processes . Food Bioprod Process . 2020 ; 123 : 1 – 13 .

Woolley E , Wanjeri A , Simeone A.   Enhancement of clean-in-place procedures in powder production using ultraviolet-induced fluorescence . Procedia CIRP . 2018 ; 78 : 364 – 369 .

Praveen U , Farnaz G , Hatim G.   Inventory management and cost reduction of supply chain processes using AI based time-series forecasting and ANN modeling . Procedia Manuf . 2019 ; 38 : 256 – 263 .

Iqbal J , Khan ZH , Khalid A.   Prospects of robotics in food industry . Food Sci Technol.   2017 ; 37 : 159 – 165 .

Kuntagod N , Podder S , Abbabathula SSS , et al.  A framework for enabling safe and resilient food factories for the public feeding programs [abstract]. In: Proceedings of the 2021 IEEE/ACM 3rd International Workshop on Software Engineering for Healthcare . Madrid, Spain; 2021 ;2103.04811. 10.1109/SEH52539.2021.00008

Naik S , Patel B.   Machine vision based fruit classification and grading—a review . Int J Comput Appl.   2017 ; 170 : 22 – 34 .

Thomas A , Sathian CT.   Cleaning-in-place (CIP) system in dairy plant—review . IOSR J Environ Sci Toxicol Food Technol.   2014 ; 8 : 41 – 44 .

Simeone A , Watson N , Sterritt I , et al.    A multi-sensor approach for fouling level assessment in clean-in-place processes . Procedia CIRP . 2016 ; 55 : 134 – 139 .

Riahi Y , Saikouk T , Gunasekaran A , et al.    Artificial intelligence applications in supply chain: a descriptive bibliometric analysis and future research directions . Expert Syst Appl . 2021 ; 173 : 114702 .

Attaran M.   Digital technology enablers and their implications for supply chain management . Supply Chain Forum . 2020 ; 21 : 158 – 172 .

Lemay DG , Baldiviez LM , Chin EL , et al.    Technician-scored stool consistency spans the full range of the Bristol scale in a healthy US population and differs by diet and chronic stress load . J Nutr.   2021 ; 151 : 1443 – 1452 .

Chin EL , Van Loan M , Spearman SS , et al.    Machine learning identifies stool pH as a predictor of bone mineral density in healthy multiethnic US adults . J Nutr.   2021 ; 151 : 3379 – 3390 .

COVID-19: a new lens for non-communicable diseases [published correction appears in Lancet . 2020;396:818] . Lancet . 2020 ; 396 : 649 . [10.1016/S0140-6736(20)31856-0]

Pinti M , Appay V , Campisi J , et al.    Aging of the immune system: focus on inflammation and vaccination . Eur J Immunol.   2016 ; 46 : 2286 – 2301 .

Rein D , Ternes P , Demin R , et al.    Artificial intelligence identified peptides modulate inflammation in healthy adults . Food Funct.   2019 ; 10 : 6030 – 6041 .

Chin EL , Simmons G , Bouzid YY , et al.    Nutrient estimation from 24-hour food recalls using machine learning and database mapping: a case study with lactose . Nutrients . 2019 ; 11 : 3045 .

Burr AHP , Bhattacharjee A , Hand TW.   Nutritional modulation of the microbiome and immune response . J Immunol.   2020 ; 205 : 1479 – 1487 .

Mirza B , Wang W , Wang J , et al.    Machine learning and integrative analysis of biomedical big data . Genes . 2019 ; 10 : 87 .

Bojar D , Powers RK , Camacho DM , et al.    Deep-learning resources for studying glycan-mediated host-microbe interactions . Cell Host Microbe.   2021 ; 29 : 132 – 144.e3 .

Fazzino TL , Martin CK , Forbush K.   The remote food photography method and SmartIntake app for the assessment of alcohol use in young adults: feasibility study and comparison to standard assessment methodology . JMIR Mhealth Uhealth.   2018 ; 6 : e10460 .

Mezgec S , Seljak BK.   Deep neural networks for image-based dietary assessment . J Vis Exp . 2021 ; e61906 . [10.3791/61906] [PMID:33779595]

Lu Y , Stathopoulou T , Vasiloglou MF , et al.    goFOOD TM : an artificial intelligence system for dietary assessment . Sensors . 2020 ; 20 : 4283 .

Boushey CJ , Spoden M , Delp EJ , et al.    Reported energy intake accuracy compared to doubly labeled water and usability of the mobile food record among community dwelling adults . Nutrients . 2017 ; 9 : 312 .

Bhat MI , Kapila R.   Dietary metabolites derived from gut microbiota: critical modulators of epigenetic changes in mammals . Nutr Rev.   2017 ; 75 : 374 – 389 .

Garrett WS.   Immune recognition of microbial metabolites . Nat Rev Immunol.   2020 ; 20 : 91 – 92 .

Cammarota G , Ianiro G , Ahern A , et al.    Gut microbiome, big data and machine learning to promote precision medicine for cancer . Nat Rev Gastroenterol Hepatol.   2020 ; 17 : 635 – 648 .

Westfall S , Carracci F , Estill M , et al.    Optimization of probiotic therapeutics using machine learning in an artificial human gastrointestinal tract . Sci Rep.   2021 ; 11 : 1067 .

Volkova A , Ruggles KV.   Predictive metagenomic analysis of autoimmune disease identifies robust autoimmunity and disease specific microbial signatures . Front Microbiol.   2021 ; 12 : 621310 .

Shima H , Masuda S , Date Y , et al.    Exploring the impact of food on the gut ecosystem based on the combination of machine learning and network visualization . Nutrients . 2017 ; 9 : 1307 .

Zeevi D , Korem T , Zmora N , et al.    Personalized nutrition by prediction of glycemic responses . Cell . 2015 ; 163 : 1079 – 1094 .

Sak J , Suchodolska M.   Artificial intelligence in nutrients science research: a review . Nutrients . 2021 ; 13 : 322 .

Doherty A , Wall A , Khaldi N , et al.    Artificial intelligence in functional food ingredient discovery and characterisation: a focus on bioactive plant and food peptides . Front Genet.   2021 ; 12 : 768979 .

Carpio LE , Sanz Y , Gozalbes R , et al.    Computational strategies for the discovery of biological functions of health foods, nutraceuticals and cosmeceuticals: a review . Mol Divers.   2021 ; 25 : 1425 – 1438 .

Shinn LM , Li Y , Mansharamani A , et al.    Fecal bacteria as biomarkers for predicting food intake in healthy adults . J Nutr.   2021 ; 151 : 423 – 433 .

Matusheski NV , Caffrey A , Christensen L , et al.    Diets, nutrients, genes and the microbiome: recent advances in personalised nutrition . Br J Nutr.   2021 ; 126 : 1489 – 1489 . doi: 10.1017/S0007114521000374

Tang W , Chen J , Wang Z , et al.    Deep learning for predicting toxicity of chemicals: a mini review . J Environ Sci Health C Environ Carcinog Ecotoxicol Rev.   2018 ; 36 : 252 – 271 .

Uesawa Y.   AI-based QSAR modeling for prediction of active compounds in MIE/AOP . Yakugaku Zasshi.   2020 ; 140 : 499 – 505 .

Chakravarti SK , Alla SRM.   Descriptor free QSAR modeling using deep learning with long short-term memory neural neural networks . Front Artif Intell.   2019 ; 2 : 17 .

Gini G , Zanoli F , Gamba A , et al.    Could deep learning in neural networks improve the QSAR models?   SAR QSAR Environ Res.   2019 ; 30 : 617 – 642 .

Miyazawa T , Itaya M , Burdeos GC , et al.    A critical review of the use of surfactant-coated nanoparticles in nanomedicine and food nanotechnology . Int J Nanomedicine.   2021 ; 16 : 3937 – 3999 .

Kozawa S , Sagawa F , Endo S , et al.    Predicting human clinical outcomes using mouse multi-organ transcriptome . iScience . 2020 ; 23 : 100791 .

Author notes

• artificial intelligence
• engineering
• science of nutrition
• pharmacy (field)
• dietary assessment
• intestinal bacteria
• toxic effect
• machine learning

Email alerts

Citing articles via.

• Recommend to your Library

Affiliations

• Online ISSN 1753-4887
• Print ISSN 0029-6643
• Copyright © 2024 International Life Sciences Institute
• About Oxford Academic
• Publish journals with us
• University press partners
• What we publish
• New features  
• Open access
• Institutional account management
• Rights and permissions
• Get help with access
• Accessibility
• Advertising
• Media enquiries
• Oxford University Press
• Oxford Languages
• University of Oxford

Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide

• Copyright © 2024 Oxford University Press
• Cookie settings
• Cookie policy
• Privacy policy
• Legal notice

This Feature Is Available To Subscribers Only

Sign In or Create an Account

This PDF is available to Subscribers Only

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

Journal of Food Science and Technology

The Journal of Food Science and Technology (JFST) is the official publication of the Association of Food Scientists and Technologists of India (AFSTI). This monthly publishes peer-reviewed research papers and reviews in all branches of science, technology, packaging and engineering of foods and food products. Special emphasis is given to fundamental and applied research findings that have potential for enhancing product quality, extend shelf life of fresh and processed food products and improve process efficiency. Critical reviews on new perspectives in food handling and processing, innovative and emerging technologies and trends and future research in food products and food industry byproducts are also welcome. The journal also publishes book reviews relevant to all aspects of food science, technology and engineering.

Due to the overwhelming number of manuscripts coming to the journal, there could be a delay of 30 days in the first decision.

Latest issue

Volume 61, Issue 6

Latest articles

Improving the oxidative stability of butter oil with nanoencapsulated ferulago angulata essential oil during accelerated shelf-life storage.

• Nassim Shavisi

Using food to color food: photostability of canistel pulp and its application as a food dye

• Fernando Antônio Anjo
• Camilla Yara Langer Ogawa
• Paula Toshimi Matumoto-Pintro

Effect of microencapsulation on physical properties of powder developed from blended oils rich in PUFA

• Neha Naijo Areekal
• A. Chakkaravarthi
• Sukumar Debnath

Exploring the preservation efficiency of cured betel leaf essential oil in augmenting the quality of fruit juice: a comprehensive evaluation of physicochemical, microbial, and sensory parameters

• Fathima Beevi Abdul Kareem
• Arunkumar Elumalai
• Baskaran Nagarethinam

Effects of the extrusion conditions, the addition of oil and the food matrix on the physical and sensory characteristics of pre-extrusion flavored products

• Isadora Giacomini Flôres
• Christian Salles
• Ana Carolina Conti

Journal updates

Biotechnology for resource efficiency, energy, environment, chemicals and health (breeech-2021), special issue 2024: international conference on new horizons in bioengineering: fostering academia-industry partnership (icb-2024).

Guest Editors 

Dr. G. Nagamaniammai,   B.E, M. Tech, Ph. D Associate Professor Department Food Process Engineering School of Bioengineering, College of Engineering and Technology SRM Institute of Science and Technology Kattankulathur- 603203 Chengalpattu (Dt) Tamil Nadu, India  Mail.Id:[email protected] and [email protected] 

Dr. P. Gurumoorthi Ph.D., FISAB. HoD (i/c), Department of Food and Process Engineering Faculty of Engineering and Technology SRM Institute of Science and Technology Kattankulathur- -603 203, Chengalpattu Dt. Tamil Nadu

Acceptance of Papers from Conference: 25th Feb 2024 to 30th Apr 2024

Journal information

• Biological Abstracts
• CAB Abstracts
• Chemical Abstracts Service (CAS)
• Current Contents/Agriculture, Biology & Environmental Sciences
• EI Compendex
• Google Scholar
• IFIS Publishing
• INIS Atomindex
• Japanese Science and Technology Agency (JST)
• OCLC WorldCat Discovery Service
• PubMedCentral
• Science Citation Index Expanded (SCIE)
• Semantic Scholar
• TD Net Discovery Service
• UGC-CARE List (India)

Rights and permissions

Springer policies

© Association of Food Scientists & Technologists (India)

• Find a journal
• Publish with us
• Track your research

Artificial intelligence in food science and nutrition: a narrative review

Affiliations.

• 1 New Industry Creation Hatchery Center (NICHe), Tohoku University, Sendai, Miyagi, Japan.
• 2 Graduate School of Agricultural Science, Tohoku University.
• 3 Graduate School of Agricultural Science, Tohoku University, Sendai, Miyagi, Japan.
• 4 Central R&D Laboratory, Kobayashi Pharmaceutical Co Ltd, Osaka, Japan.
• 5 International Life Sciences Institute (ILSI) Japan, Tokyo, Japan.
• 6 Co-Creation Center, Meiji Holdings Co Ltd, Tokyo, Japan.
• 7 Milk Science Research Institute, Megmilk Snow Brand Co Ltd, Saitama, Saitama, Japan.
• 8 Health & Wellness Products Research Laboratories, Kao Corp, Tokyo, Japan.
• 9 Research Institute for Creating the Future, Fuji Oil Holdings Inc, Tsukubamirai, Ibaraki, Japan.
• 10 Health Science Research Center, Morinaga & Co Ltd, Yokohama, Kanagawa, Japan.
• 11 Food Ingredients & Technology Institute, R&D Division, Morinaga Milk Industry Co Ltd, Zama, Kanagawa, Japan.
• PMID: 35640275
• DOI: 10.1093/nutrit/nuac033

In the late 2010s, artificial intelligence (AI) technologies became complementary to the research areas of food science and nutrition. This review aims to summarize these technological advances by systematically describing the following: the use of AI in other fields (eg, engineering, pharmacy, and medicine); the history of AI in relation to food science and nutrition; the AI technologies currently used in the agricultural and food industries; and some of the important applications of AI in areas such as immunity-boosting foods, dietary assessment, gut microbiome profile analysis, and toxicity prediction of food ingredients. These applications are likely to be in great demand in the near future. This review can provide a starting point for brainstorming and for generating new AI applications in food science and nutrition that have yet to be imagined.

Keywords: artificial intelligence; deep learning; dietary assessment; food components; gut microbiome; health; immunological functions; machine learning; timeline; toxicity prediction.

© The Author(s) 2022. Published by Oxford University Press on behalf of the International Life Sciences Institute. All rights reserved. For permissions, please e-mail: [email protected].

Publication types

• Research Support, Non-U.S. Gov't
• Artificial Intelligence*
• Delivery of Health Care*
• Food Technology

food science Recently Published Documents

Total documents.

• Latest Documents
• Most Cited Documents
• Contributed Authors
• Related Sources
• Related Keywords

Stimulus-responsive hydrogels in food science: A review

Dietary fiber in plant cell walls—the healthy carbohydrates.

Abstract Dietary fiber (DF) is one of the major classes of nutrients for humans. It is widely distributed in the edible parts of natural plants, with the cell wall being the main DF-containing structure. The DF content varies significantly in different plant species and organs, and the processing procedure can have a dramatic effect on the DF composition of plant-based foods. Given the considerable nutritional value of DF, a deeper understanding of DF in food plants, including its composition and biosynthesis, is fundamental to the establishment of a daily intake reference of DF and is also critical to molecular breeding programs for modifying DF content. In the past decades, plant cell wall biology has seen dramatic progress, and such knowledge is of great potential to be translated into DF-related food science research and may provide future research directions for improving the health benefits of food crops. In this review, to spark interdisciplinary discussions between food science researchers and plant cell wall biologists, we focus on a specific category of DF—cell wall carbohydrates. We first summarize the content and composition of carbohydrate DF in various plant-based foods, and then discuss the structure and biosynthesis mechanism of each carbohydrate DF category, in particular the respective biosynthetic enzymes. Health impacts of DF are highlighted, and finally, future directions of DF research are also briefly outlined.

Agricultural and Food Science: Historic Growth in Breadth and Impact

Sociobiodiversidade e soberania e segurança alimentar e nutricional como um direito indissociável à alimentação adequada e saudável.

A biodiversidade é fundamental para uma vida digna e justa. Assim, o presente trabalho teve como objetivo contextualizar a sociobiodiversidade e Soberania e Segurança Alimentar e Nutricional (SSAN) como um direito indissociável à alimentação adequada e saudável. As buscas foram realizadas no Portal de periódicos CAPES, FSTA – Food Science and Technology Science Direct, SciELO, Google Acadêmico, literatura cinzenta e portal de Organização das Nações Unidas (UN), Ministério do Meio Ambiente - MMA, Biodiversity for Food and Nutrition – BFN e legislações especificas. A revisão mostrou a importância de estratégias de sistemas alimentares saudáveis e sustentáveis, mediante ações transversais em políticas de governança para mitigar a injustiça social, impulsionar a econômica dos povos, promover ações de biodiversidade, sustentabilidade e SSAN como direito à alimentação adequada e saudável. Essas ações promovem a geração de emprego, renda e melhor qualidade de vida das famílias de agricultores, às comunidades tradicionais, produtores agroecológicos e assentados da reforma agrária que tem a terra como seu sustento. Além disso, promovem o meio ambiente sustentável como meta para ODS 2030, contribuindo para a humanidade enfrentar os desafios, em âmbito global, da injustiça social como direito à alimentação adequada e saudável.

The Influence of the Western Diet on Microbiota and Gastrointestinal Immunity

Diet exerts a major influence upon host immune function and the gastrointestinal microbiota. Although components of the human diet (including carbohydrates, fats, and proteins) are essential sources of nutrition for the host, they also influence immune function directly through interaction with innate and cell-mediated immune regulatory mechanisms. Regulation of the microbiota community structure also provides a mechanism by which food components influence host immune regulatory processes. Here, we consider the complex interplay between components of the modern (Western) diet, the microbiota, and host immunity in the context of obesity and metabolic disease, inflammatory bowel disease, and infection. Expected final online publication date for the Annual Review of Food Science and Technology, Volume 13 is March 2022. Please see http://www.annualreviews.org/page/journal/pubdates for revised estimates.

Innovative Treatments Enhancing the Functionality of Gut Microbiota to Improve Quality and Microbiological Safety of Foods of Animal Origin

The gastrointestinal tract, or gut, microbiota is a microbial community containing a variety of microorganisms colonizing throughout the gut that plays a crucial role in animal health, growth performance, and welfare. The gut microbiota is closely associated with the quality and microbiological safety of foods and food products originating from animals. The gut microbiota of the host can be modulated and enhanced in ways that improve the quality and safety of foods of animal origin. Probiotics—also known as direct-fed microbials—competitive exclusion cultures, prebiotics, and synbiotics have been utilized to achieve this goal. Reducing foodborne pathogen colonization in the gut prior to slaughter and enhancing the chemical, nutritional, or sensory characteristics of foods (e.g., meat, milk, and eggs) are two of many positive outcomes derived from the use of these competitive enhancement–based treatments in food-producing animals. Expected final online publication date for the Annual Review of Food Science and Technology, Volume 13 is March 2022. Please see http://www.annualreviews.org/page/journal/pubdates for revised estimates.

Trend and status of Food Science and Technology category based on the Essential Science Indicators during 2011 – 2021

Plant-based proteins: the good, bad, and ugly.

Our global population is growing at a pace to exceed 10 billion people by the year 2050. This growth will place pressure on the agricultural production of food to feed the hungry masses. One category that will be strained is protein. Per capita protein consumption is rising in virtually every country for both nutritional reasons and consumption enjoyment. The United Nations estimates protein demand will double by 2050, and this will result in a critical overall protein shortage if drastic changes are not made in the years preceding these changes. Therefore, the world is in the midst of identifying technological breakthroughs to make protein more readily available and sustainable for future generations. One protein sourcing category that has grown in the past decade is plant-based proteins, which seem to fit criteria established by discerning consumers, including healthy, sustainable, ethical, and relatively inexpensive. Although demand for plant-based protein continues to increase, these proteins are challenging to utilize in novel food formulations. Expected final online publication date for the Annual Review of Food Science and Technology, Volume 13 is March 2022. Please see http://www.annualreviews.org/page/journal/pubdates for revised estimates.

Microbial nanotechnology: New horizons in food science and technology

Introductory microbiology lab skills and techniques in food science, export citation format, share document.

EDITORIAL article

This article is part of the research topic.

Climate Science, Solutions and Services for Net Zero, Climate-Resilient Food Systems

Editorial: Climate Science, Solutions and Services for Net Zero, Climate-Resilient Food Systems Provisionally Accepted

• 1 Hadley Centre, Met Office, United Kingdom
• 2 School of Biological Sciences, University of Bristol, United Kingdom
• 3 Global Sustainability Institute, Anglia Ruskin University, United Kingdom
• 4 Wageningen Economic Research, Wageningen University and Research, Netherlands
• 5 School of Biology, Faculty of Biological Sciences, University of Leeds, United Kingdom
• 6 The James Hutton Institute, United Kingdom

The final, formatted version of the article will be published soon.

Food systems are both a major contributor to global greenhouse gas emissions (Costa et al., 2022) and strongly impacted by climate and weather (Falloon et al., 2022). Solutions to deliver net zero food systems therefore need to take climate impacts, adaptation, and resilience into account to ensure they are appropriate in a changing climate and do not conflict with adaptation goals. Food system adaptation options must also consider potential trade-offs, consequences, and synergies with net zero and other objectives such as the Sustainable Development Goals. Solutions for net zero, climate resilient food systems therefore require systematic, interdisciplinary approaches across academia, governments, business, NGOs, and the public. This Research Topic showcases a collection of studies covering cutting edge science and thought leadership towards the goal of net zero, climate-resilient food systems.Several papers use case study events or assess current and future practical climate adaptation and net zero practices in farming and food systems businesses. An exploration of farmers' perception of climate adaptation strategies in the rice-growing zone of Punjab, Pakistan (Khan et al., 2022) revealed significant perceived climate changes, while the extent of adaptation was strongly linked to education and access to climate information and credit services. The principal factors determining adaptation decisions included farmers' age, primary occupation, income, landholding, access to irrigation, credit, climate information, and agricultural extension services; hence improving the alterable factors amongst these should improve resilience of the rice farming system. Sarker et al. (2022) assess the benefits of conservation tillage and residue management to soil health and crop productivity in a Bangladeshi rice-maize cropping system. Compared to conventional tillage, the overall improvement in soil conditions gradually increased crop productivity, and improved farm profitability compared to conventionally tilled rice and maize crops. Conservation agriculture could therefore be an appropriate practice for sustaining soil fertility and crop yield under rice-maize systems in light-textured soils in Bangladesh.Kumar Jha et al. (2023) report results from experimental studies in the rice-wheat system of Bihar, India that evaluate the feasibility of early rice transplanting combined with a community irrigation approach. These practices increased rice yield and water productivity, compared to late-sown crops, while timely wheat harvesting allowed cultivation of an additional summer crop. Overall, this approach to managing climatic risks and variability increased the productivity of the rice-wheat cropping system. Sakrabani (2024) analyses the opportunities and challenges for organo-mineral fertilisers (OMFs) in enabling food security and meeting net zero goals, identifying policy interventions that balance environmental protection and meeting food security. Short-term priorities include development of guidelines, energy incentives for drying feedstocks and renewable energy; in the medium-term, evidence gathering from long-term field trials, funding to support innovation, and regional policy harmonisation; and in the long-term feedstock certification and joined-up waste-fertilizer policy. 2023) use the record-breaking United Kingdom heatwave of 2022 as a case study to explore the impacts on the poultry and wheat sectors, and to identify potential adaptation options for a climate-resilient, net-zero food system. Both negative and positive heatwave impacts were felt across the food system, from greater energy costs for cold storage, retail refrigeration failure, and livestock heat stress but also increased wheat yields. A range of adaptation measures are proposed for both poultry and wheat. 2023) present a novel methodology for developing a sustainable business model (SBM) in the food, beverage, and tobacco sector, using data from 252 businesses that reported to the Carbon Disclosure Project (CDP). Their analysis identified, prioritized and mapped a range of environmental sustainability themes and 150 green practices that could contribute to emission reduction targets, resulting in a net-zero value proposition to customers. The remaining papers tackle key challenges at the broader policy level. Gelardi et al. ( 2023) review the evidence for agricultural soils to contribute to net zero goals, examine existing support strategies and emerging markets, and recommend ways to synthesize approaches into a cohesive policy portfolio for the US to deliver effective and equitable outcomes. 2023) apply a multi-level participatory scenario approach combined with modelling and decision support tools to develop scenarios in support of future food security policy in Bangladesh. Their future scenarios show that diverse pathways are possible, but with very different food security and low-carbon development outcomes. Andrews et al. (2023) draw on agroecological principles to propose a framework for aligning foodsystems policy to provide multiple benefits. Their six-part framework can underpin public health, environmental sustainability, economic stability, social cohesion, and national security and sovereignty. The seven tactical implementation principles they propose can help integrate community-scale efforts to establish food systems and ensure food systems policy effectiveness. To advance solutions and services that support the goal of climate-resilient, net-zero food systems and better food security outcomes, several key themes emerge from the papers presented here, noting that the challenges highlighted below should not dissuade action (Gelardi et al., 2023):1. Broad and diverse stakeholder engagement across the agri-food supply-chain and beyond in solution co-design and development (Asif et al., 2023;Gelardi et al., 2023), including youth and poor rural communities (Moghayer et al., 2023). 2. Effective integration and joint prioritization of climate adaptation and mitigation options, alongside consideration of their trade-offs, consequences and co-benefits and interactions. This should include social, economic and environmental dimensions and pressures for land (Davie et al., 2023;Gelardi et al., 2023;Kumar Jha et al., 2023), and balancing short and long-term priorities (Moghayer et al., 2023). 3. Addressing barriers to adoption and structural issues (Davie et al., 2023;Gelardi et al., 2023;Khan et al., 2022) in climate adaptation and net-zero. 4. Integrated policy that supports effective environmental stewardship and is underpinned by well-functioning governance systems and political will (Andrews et al., 2023;Moghayer et al., 2023;Sakrabani, 2024). 5. Enabling shifts in consumer behaviour (Moghayer et al., 2023). 6. Implementation of practice-and place-specific programs of change (Gelardi et al., 2023)

Keywords: food systems, Climate resilience, Climate Change, net zero, Solutions, Climate services, Weather and climate, climate impacts

Received: 12 Apr 2024; Accepted: 16 Apr 2024.

Copyright: © 2024 Falloon, Jones, Van Berkum, Kepinski and Rivington. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

* Correspondence: Dr. Pete Falloon, Met Office, Hadley Centre, Exeter, EX1 3PB, Devon, United Kingdom

People also looked at

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts

Latest science news, discoveries and analysis

Could a rare mutation that causes dwarfism also slow ageing?

Bird flu in US cows: is the milk supply safe?

Future of Humanity Institute shuts: what's next for ‘deep future’ research?

Judge dismisses superconductivity physicist’s lawsuit against university

Nih pay raise for postdocs and phd students could have us ripple effect, hello puffins, goodbye belugas: changing arctic fjord hints at our climate future, china's moon atlas is the most detailed ever made, ‘shut up and calculate’: how einstein lost the battle to explain quantum reality, ecologists: don’t lose touch with the joy of fieldwork chris mantegna.

Should the Maldives be creating new land?

Lethal AI weapons are here: how can we control them?

Algorithm ranks peer reviewers by reputation — but critics warn of bias

How gliding marsupials got their ‘wings’

Bird flu virus has been spreading in us cows for months, rna reveals, audio long read: why loneliness is bad for your health, nato is boosting ai and climate research as scientific diplomacy remains on ice, rat neurons repair mouse brains — and restore sense of smell.

Retractions are part of science, but misconduct isn’t — lessons from a superconductivity lab

Any plan to make smoking obsolete is the right step

Citizenship privilege harms science

European ruling linking climate change to human rights could be a game changer — here’s how charlotte e. blattner, will ai accelerate or delay the race to net-zero emissions, current issue.

The Maldives is racing to create new land. Why are so many people concerned?

Surprise hybrid origins of a butterfly species, stripped-envelope supernova light curves argue for central engine activity, optical clocks at sea, research analysis.

Ancient DNA traces family lines and political shifts in the Avar empire

A chemical method for selective labelling of the key amino acid tryptophan

Robust optical clocks promise stable timing in a portable package

Targeting RNA opens therapeutic avenues for Timothy syndrome

Bioengineered ‘mini-colons’ shed light on cancer progression, galaxy found napping in the primordial universe, tumours form without genetic mutations, marsupial genomes reveal how a skin membrane for gliding evolved.

Scientists urged to collect royalties from the ‘magic money tree’

Breaking ice, and helicopter drops: winning photos of working scientists

Shrouded in secrecy: how science is harmed by the bullying and harassment rumour mill

Want to make a difference try working at an environmental non-profit organization, how ground glass might save crops from drought on a caribbean island, books & culture.

How volcanoes shaped our planet — and why we need to be ready for the next big eruption

Dogwhistles, drilling and the roots of Western civilization: Books in brief

Cosmic rentals

Las borinqueñas remembers the forgotten puerto rican women who tested the first pill, dad always mows on summer saturday mornings, nature podcast.

Latest videos

Nature briefing.

An essential round-up of science news, opinion and analysis, delivered to your inbox every weekday.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

South Africa’s young education researchers need networks to share experience more than pressure to produce outputs

• Loadshedding
• South African News
• African News
• International News
• All Opinions
• Latest Opinions
• Raymond Suttner - Suttner's View
• Politics 360 - Ebrahim Fakir
• Institute for Security Studies
• South African Institute of International Affairs
• Saliem Fakir - Low Carbon Future
• The Conversation
• Real Economy
• Other Opinions
• Author Interviews
• All Legislation
• Constitution
• Audio Articles
• Recommendations
• All Case Law
• Constitutional Court
• Supreme Court of Appeal
• High Courts
• All Legal Briefs
• Company Posts
• Legal Notice

Note: Search is limited to the most recent 250 articles. To access earlier articles, click Advanced Search and set an earlier date range. To search for a term containing the '&' symbol, click Advanced Search and use the 'search headings' and/or 'in first paragraph' options.

And must exclude these words...

Sponsored by

Please enter the email address that you used to register on Polity.org.za. Your password will be sent to this address.

Email this article

separate emails by commas, maximum limit of 4 addresses

Article Enquiry

South africa’s young education researchers need networks to share experience more than pressure to produce outputs, embed video.

Embed Video Popup Video Instagram

Download Buy Photos

26th April 2024

ARTICLE ENQUIRY       SAVE THIS ARTICLE       EMAIL THIS ARTICLE

Font size: - +

Good research matters. It can have broad, positive consequences. On paper South Africa’s government recognises this. The 2019 White Paper on Science, Technology and Innovation points out how knowledge from many disciplines interacts to deepen awareness of South Africa’s serious and long-standing challenges, like energy, food security and inequality. More importantly, this knowledge can be used to tackle these problems.

To this end, the Department of Higher Education and Training and statutory bodies like the National Research Foundation actively encourage research production. This is done through, for instance, funded research chairs and financial incentives to publish . Universities, too, are active here. They offer workshops and seminars on topics such as funding opportunities, research ethics, grant writing, and postgraduate supervision.

But several factors prevent these measures from bearing as much fruit as they could – particularly for those researchers just beginning their academic careers. In 2019 Universities South Africa held a symposium about early career researchers. It found that this cohort struggled to make time to conduct and publish research. The young researchers also wanted far more support when it came to, among other things, applying for grants and developing their research profiles.

In 2020 the South African Education Research Association established an early career researchers portfolio on its executive committee. We are both executive members of the association. We are also education researchers. We conducted a study to investigate how early career education researchers experience the organisation’s strategies.

The researchers we interviewed said they appreciated the practical skills and networking opportunities offered by the association. They valued its culture of collaboration and openness to sharing resources and knowledge.

Our findings show that there is a generation of dedicated, motivated younger education researchers in South Africa. They are asking for more training in areas like writing, plagiarism and citation, and supervising postgraduate students. They also want more networking opportunities and the chance to collaborate with each other and with older, more established scholars. This kind of support should be provided by more than just professional organisations like ours. Universities and research institutions have a crucial role to play, too.

Key findings

The South African Education Research Association defines early career researchers as those who are engaged in postgraduate study or are embarking on a research career at a higher education institution or any other research centre.

As part of our support for these researchers the association offers, among other things, doctoral awards, public seminars and mentorships. We also provide training in writing for publication. And we aim to establish a community which provides critical and supportive engagement.

For the study we approached a sample of 34 early career researchers in the educational research field. They had all taken part in the association’s various activities. Participation in the study was voluntary. We received 21 responses via an online survey. Participants were asked to describe their involvement with the association, and their positive and negative experiences of this involvement.

Many of these were positive. Said one respondent:

I enjoyed the (early career researcher) workshop, and the (association’s) conference gave me the opportunity to showcase my PhD research.

Another told us:

I have gone on to publish my presentation papers after benefiting from feedback from colleagues at conferences.

Other positive factors they highlighted included a culture of collaboration within a supportive research environment and the chance to network with more experienced researchers, as well as with their peers.

Researchers also told us how engaging with experienced scholars through the association’s programmes had positively influenced their own research trajectories.

A few participants told us their institutions were not providing enough support or proper mentoring. This made them especially grateful for the association’s approach, particularly as it related to openly sharing resources and knowledge.

They also identified some gaps in the association’s programmes.

For instance, they wanted more practical assistance in various areas of scholarship. These included topics like plagiarism and citation and the supervision of postgraduate students. Others wanted help in choosing the best design and methodology for their research.

Participants asked for more opportunities for novice researchers to engage and collaborate with seasoned academics. This was seen as especially important for those working at less research-intensive institutions.

The researchers also called for greater international collaboration, especially among Southern African Development Community (SADC) countries. It is encouraging that the organisation’s 2024 conference will be held in conjunction with the Southern African Educational Research Network. The network includes associations from Botswana, Lesotho, Eswatini and Namibia.

Finally, they wanted more avenues to be created for funding and knowledge sharing among early career researchers.

Broaden ideas of capacity-building

Our findings show the limitations of an approach to building young researchers’ skills and abilities that focuses predominantly on producing more research outputs.

This echoes the argument made by researchers Jack Lee and Aliya Kuzhabekova : an accounting of publications, patents and doctorates does not fully capture the complexity of capacity building. Instead, building research capacity involves shifting from, as researchers Alison Lee and David Boud put it , “what is produced [outputs] to the production of the person who produces”.

We are not denying that resources and infrastructure play an essential role in building research capacity. Our participants all flagged a lack of funding to attend conferences as a major hindrance. However, we would argue that attention must simultaneously be given to building cultures of collaboration and support that allow peer networks to flourish and reciprocal learning to become the norm.

Encouragingly, South Africa’s Council on Higher Education supports this kind of approach. It has called for doctoral programmes to enable students to engage with a wide range of stakeholders and communities outside their immediate research groups.

A strategy built on cultures of collaboration and support will encourage knowledge production and scholarly growth. This will help more young researchers to advance their careers. And that’s good for South Africa, as it will add to the country’s pool of experienced researchers.

Written by Mpho-Entle Puleng Modise , Associate professor, University of South Africa and Maureen Robinson , Emeritus Professor, Faculty of Education, Stellenbosch University

This article is republished from The Conversation under a Creative Commons license. Read the original article .

EMAIL THIS ARTICLE       SAVE THIS ARTICLE ARTICLE ENQUIRY

To subscribe email [email protected] or click here To advertise email [email protected] or click here

Polity.org.za is a product of Creamer Media. www.creamermedia.co.za

Other Creamer Media Products include: Engineering News Mining Weekly Research Channel Africa

Newsletters

Sign up for our FREE daily email newsletter

Subscriptions

We offer a variety of subscriptions to our Magazine, Website, PDF Reports and our photo library.

Subscriptions are available via the Creamer Media Store.

Advertising on Polity.org.za is an effective way to build and consolidate a company's profile among clients and prospective clients. Email [email protected]