[Feature: The Future of Japanese "Food"] Ryuichi Suzuki: The Power of Japanese Food Brought Out by AI Taste Sensors

2022/02/04

I serve as the representative of a company called AISSY, which has developed and commercialized a taste sensor named "Leo" that can break down and quantify the taste of any food or drink into five basic tastes (sweetness, saltiness, sourness, bitterness, and umami). AISSY is a company founded based on research results from Keio University, with investment from Keio. The company name is an abbreviation for Artificial Intelligence SensingSYstems, but it also carries a second meaning by playing on the sound of "I see," representing the visualization of the invisible (taste).

Through various analyses with Leo, I feel I have begun to see a bright future for Japan through food. Here, I would like to introduce taste sensor technology and then write about the potential of Japanese food.

The roots of Leo go back to my days as an undergraduate at Keio. When I was a sophomore, there was a ramen shop near the Hiyoshi Campus that was, frankly, not very delicious. It was the kind of place where only about five customers would come in a single day. Due to certain circumstances, I ended up helping manage that ramen shop with several Keio friends, but the discussions we had about taste at that time were very difficult to navigate. As a science student, I liked discussing things using numbers even back then, but numbers never came up in talks about taste. Instead, the conversations were entirely qualitative, such as "Let's reduce the salt a bit more" or "Let's add this topping to bring out a little more umami." At that time, I felt a potential need for the quantification of taste.

Later, as a senior, I saw my academic advisor at the time, Professor Koji Suzuki, and my seniors researching the technology that would become the basis for the current Leo, and it connected with my experience at the ramen shop. I thought that food and beverage manufacturers must be discussing taste even more than I did at the ramen shop, so there had to be customers who wanted to quantify it. Taste sensors had existed to some extent before then, but they faced challenges such as the inability to measure sweetness and low accuracy due to linear analysis methods.

The AI taste sensor developed at Keio was a groundbreaking technological seed that solved the problems of previous taste sensors. I felt it had a strong chance of success as a business, so I founded AISSY immediately after completing graduate school.

At AISSY, we conducted joint research with Keio University and developed a technology capable of sensing the five tastes humans perceive (sweetness, saltiness, sourness, bitterness, and umami). This was achieved by using an analysis method called neural network analysis, which considers taste interactions, for the component concentrations detected by electrochemical sensors. Figure 1 shows the conceptual diagram and error comparison of the taste sensor (the small error indicates that it can reproduce human taste). Neural networks are the fundamental concept behind the AI that is currently booming.

In product development, Leo is often used to determine how to get closer to a target taste. In marketing, many client companies increase sales by highlighting high levels of sweetness or umami. If you are interested, please visit this site.

As I researched taste, I began to think that "the Japanese sense of taste is miraculous," and I even published a book titled "The Japanese Sense of Taste is the Best in the World" (Kosaido Shinsho).

The greatness of the Japanese sense of taste can be summarized in one word: sensitivity to umami. As shown in Figure 2, results from Leo's research show that when comparing Japanese food on average, it has a much stronger umami profile compared to food from overseas. The true identity of that feeling of "something is missing, I miss Japanese food" when going abroad is "umami."

The reason why our food culture became so sensitive to umami is related to Japan's geographical factor as an island nation surrounded by the sea. You might think the UK is the same, but while the variety of fish caught in the UK is not that large, Japan is not only surrounded by the sea but also has a long coastline and serves as a meeting point for northern and southern currents. Thus, it is blessed with seafood and abundant marine resources.

To begin with, for a country or region to grow to a population of over 10 million, it is essential to consume animal protein. Populations increase by eating meat and fish, but because fish yields fluctuate much more wildly than meat, fish was relegated to the position of a meat substitute in other regions outside of Japan. However, during the Edo period, Japan was able to provide animal protein for a population of about 20 to 30 million almost entirely through fish. While obtaining animal protein from seafood, the culture of dashi (broth), such as bonito dashi and kelp dashi, became highly developed. This is the reason why Japanese food is rich in umami.

Additionally, Japanese people cultivated their sense of sweetness through white rice. Compared to the sweetness of sugar, white rice releases more sweetness the more you chew it, leading people to savor it more, which naturally sharpened their taste receptors.

The environment closest to Japan's was southeastern China, and in fact, the prototype of sushi was born there. Animal protein is difficult to preserve for long periods, and as a solution, fish was placed in cooked rice for preservation. The rice fermented naturally, and the lactic acid produced during that time acted to prevent the fish from spoiling. This was introduced to Japan and became sushi. However, it is said that on the continent, it was eventually phased out due to the meat-oriented thinking of ethnic groups coming from inland.

Food is a culture, so it develops and it also declines. To develop, new combinations are necessary. For example, Nikujaga (meat and potato stew), which is common today, uses Western ingredients with Japanese seasoning, so it was an impossible combination until the Meiji era. Before long, it gained public acceptance and spread.

Taste also changes with the times. Recently, Japanese people have become accustomed to the bitterness of coffee. Originally, that kind of sharp bitterness did not exist in Japan. Certainly, tea has bitterness, but while tea is bitterness + umami, coffee is bitterness + sourness, resulting in a sharp bitterness. It is easy to understand if you imagine Italian cuisine; in Western cuisine, the combination of sourness and umami is popular, whereas in Japanese food, the combination of saltiness and umami is orthodox.

Currently, taste is becoming globalized, and the tastes of different countries are not as distinct as they used to be. In recent years, the preference for umami has been increasing worldwide, and the whole world is becoming "umami-loving," which is a great opportunity for Japanese food, which is rich in umami. By exporting the rare sense of taste and food culture of the Japanese—born from an exquisite distance from the continent—to the rest of the world, the future of Japan will become brighter as Japanese food spreads further.

In doing so, it will be important to arrange it in a way that is easy for the local area to accept, just like the aforementioned Nikujaga. Humans are wary of "tastes that are too new" and get bored with "the usual taste," so a "slightly new taste" is best. I want to make it possible to discuss how to achieve a "slightly new taste" quantitatively using taste data. It is theoretically possible to quantify the taste of local cuisine, have AI learn it, and calculate what kind of taste will be accepted, so all that remains is to create success stories.

The future of realizing "product development based on scientific data, not just intuition," which I felt during my days as a Keio student challenging the reconstruction of a ramen shop, is just around the corner.