Hokkaido has the highest calorie-based food self-sufficiency rate in Japan and the fourth highest rate of production value-based food self-sufficiency. It is needless to say that Hokkaido is vitally important for Japan’s food production. Muroran Institute of Technology is collaborating with companies to conduct research on robot cultivation technology and automatic greenhouse cultivation control technology using AI (artificial intelligence), as well as research to enhance the functionality of food. Here, we interviewed two professors who are engaged in research related to food and agriculture.
Associate Professor Watanabe:I specialize in the construction of algorithms, which are procedures for optimization and data mining, in the field of information processing. I am researching the development of problem-solving procedures that consider multiple criteria, called “multi-objective optimization,” and the mining (visualization) method of results.
I also analyze and propose solutions that meet the needs of companies, such as low-cost, high-quality production environment model optimization. What I am involved in from the perspective of food and agriculture is to create models for vegetable greenhouse cultivation. A company with which Muroran Institute of Technology has a comprehensive cooperation agreement cultivates tomatoes in a glass greenhouse the size of several gymnasiums as part of its business.
Concerning tomatoes shipped to supermarkets, it is necessary to make an accurate yield prediction because if the actual yield is more or less than the quantity for delivery in the contract, the producer will incur a loss. There is also an issue of how to accurately observe the situation in the vast greenhouse and how to properly engage in cultivation according to the situation. In the joint research, we are working on such yield prediction, state estimation, and cultivation management utilizing AI and IoT technologies.
Associate Professor Tokuraku:My main research theme is protein aggregation, which causes various diseases. For example, the cause of Alzheimer’s disease is the aggregation of proteins such as amyloid β and tau in the brain, which causes nerve cell death. How do these proteins aggregate and cause nerve cells to die? And how can this aggregation be suppressed? I am studying these using the bio-nano imaging technology I developed and AI.
Furthermore, I am collaborating with some companies to analyze foods that cause lifestyle-related diseases and to develop foods that have the functionality to suppress specific diseases. Professor Watanabe is engaged in a tomato production model, and I am working to improve the functionality of tomatoes to increase their added value. Although we are involved in different projects, we hold regular meetings and collaborate in research.
Associate Professor Watanabe:Although Hokkaido is known as an agricultural kingdom, the average age of producers is in their 60s. Considering 10 or 20 years from now, it is an urgent task to make an environment where even a small number of people can stably harvest agricultural products. In that sense, I believe that research on greenhouse cultivation, which has a stable growing environment compared to open-field cultivation, is an important project that will support agriculture in Hokkaido in the future. Vegetable harvesting has relied on the intuition and experience of the growers who are in charge of cultivation. We are analyzing their know-how using AI and making AI learn and optimize the adjustment of temperature, humidity, sunlight, and fertilizer to build a production model that allows even a small number of people to harvest high-quality crops with a high yield.
Associate Professor Tokuraku:I think that maintaining health by ingesting products processed from highly functional crops will be an area of increasing concern in the future society. Not to mention the good yield and quality, adding value to products by improving their functionality, which leads to health, is also important to increase the value of agriculture in Hokkaido.
Growers, who are professionals in cultivation, control the growth of tomatoes by changing the amount of fertilizer required depending on the condition of the tomatoes. That is Professor Watanabe’s research field. I am interested in what kind of functional substances tomatoes produce when loads are placed on the cultivation environments. Although functionality of foods has a wide range of effects, by utilizing AI for its analysis, the speed of development of agricultural products with the functionality that companies aim for will accelerate further in the future.
Associate Professor Watanabe:In my research field, with the development of devices and IoT technology, it has become easier to perform quick measurement and analysis at low cost. It took time and was costly in the past even if AI was used. In the field of AI technology, deep learning, in which machines automatically learn and derive answers, is now attracting attention, but there are still issues to be solved, such as being unable to explain why the answers were derived.
In the future, it will be important to develop a machine learning model called “X (explainable) AI” that can explain the process leading to prediction and estimation results to humans, and I would like to work on its development.
Associate Professor Tokuraku:As Professor Watanabe says, the development of technologies such as AI has made it possible to visualize things that were difficult to quantify, making it easier to identify the cause of disease. In research on Alzheimer’s disease, joint research with companies have revealed that Labiatae plants used for herbal spices have the effect of suppressing the aggregation of amyloid β. We are also expanding the range of applications to Parkinson's disease, rheumatism, diabetes, and other diseases where protein aggregation is said to be involved in the onset as in Alzheimer's disease.
In the future, I would like to further develop visualization technology, strive to identify the cause of diseases, discover effective foods grown in Hokkaido, and conduct research to further enhance functionality.