HACARUS Brings Sparse Modeling to Drug Discovery with Mitsubishi Tanabe Pharma

HACARUS Brings Sparse Modeling To Drug Discovery With Mitsubishi Tanabe Pharma

Jointly developed screening model leads to faster analysis and feature extraction capabilities

May 25, 2021 – Kyoto, Japan – HACARUS, the provider of big insights from small data, today announced the strong results of joint research with Mitsubishi Tanabe Pharma Corporation related to the screening process, focused on first-in-class drug discovery. Utilizing HACARUS’ proprietary AI technology, the new model drastically reduces processing time and solves “black box” issues related to the feature extraction for hit compounds.

“From this project, we were able to develop some very powerful analytical techniques, making the analysis process 56 times more efficient than before. Continuing with our joint research, I look forward to improving this model further with HACARUS and see how we can implement these results in actual drug discovery projects.” said a medical researcher from Mitsubishi Tanabe Pharma Corporation.

The use of HACARUS’ sparse modeling based AI, as opposed to conventional machine learning methods, provided several advantages that have contributed to the project’s success. The approach allows the use of a common model for multiple compounds, drastically shortening the computational time from 20-40 minutes down to around 16 seconds. Additionally, by combining the Kullback-Leibler divergence method with Sparse structure learning, it has become possible to determine which feature correlations have changed and to what extent bringing crucial explanablity into the field.

By using this HACARUS unique model, it may become possible to screen a variety of conditions including disease type and genetic polymorphism based on a patient’s iPS cells. Going forward, HACARUS’ Sparse modeling technology will play an integral role in developing the methods used to realize these possibilities.



HACARUS INC, founded 2014 in Kyoto, Japan is the leading provider of Explainable Lightweight AI Tools, backed by Osaka Gas and Miyako Capital (Kyoto University) among others. Its solutions are used in the Medical field to enable humans to make better, faster and more reliable decisions, based on AI driven insights. HACARUS’ proprietary AI engine is built using Sparse Modeling, an AI method that understands data like a human would – by its unique key features and is far more resource, time and energy efficient when compared to Deep Learning. To learn more, visit https://hacarus.com.


About Mitsubishi Tanabe Pharma Corp.

Mitsubishi Tanabe Pharma Corporation is committed to improving the lives of people around the world through the creation of pharmaceuticals, with focus on topics relating to the central nervous system, immune-inflammation, diabetes and kidneys, and vaccines. The core of their work lies with the trust of their patients, healthcare professionals, and society. To learn more, visit https://www.mt-pharma.co.jp/e/.


Takashi Someda

CTO of Hacarus. Has over 15 years experience as a software and server engineer in several global SaaS startups. Currently working hard to make something new with machine learning. Holds masters degree in Information Science at Kyoto University.

Subscribe to our newsletter

Click here to sign up