Japan's Kirin taps AI for beer making
Machine learning expected to yield new recipes
TOKYO -- Kirin Brewery will employ artificial intelligence to perform beer-making functions usually entrusted to human professionals in a scheme to shorten product development and speed training.
The Kirin Holdings unit is pairing with the Mitsubishi Research Institute to install AI brew test programs as early as this year. The applications will determine the desired flavor, aroma, color and alcohol content, then produce the corresponding recipe.
The machine-learning program will discover optimum brewing formulas based on 20 years' worth of test data. It takes a professional brewer a decade or more to hone such skills.
Not only is the industrial landscape changing fast, Kirin also faces an increasingly fragmentary beer market in terms of preferences. The Japanese company seeks to maintain its competitive edge by developing new offerings in an efficient manner.
The use of AI in beer production will also streamline the transfer of expertise to new hires. "AI could discover more efficient methods," said a Kirin representative.
Meanwhile, Ajinomoto is also planning to leverage AI to completely automate the process of fermenting amino acids. The Tokyo-based food producer will quantify the conditions that lend to productive fermentation by fiscal 2019. Ajinomoto will create a data-sharing system that will boost cost competitiveness internationally.