top of page

SOLUTIONS
​ Solution

Epistra is a technology company that develops technology to accelerate research and development with AI and robots.

Our solutions ​ are mainly used in the life sciences.

Solution 01

EpistraAccelerate

with state-of-the-art machine learning technology

Shorten R&D period and improve yield

In life science research and development, there are various challenges due to repeated manual experiments and trial and error. Epistra Accelerate, an automatic experiment optimization system using AI, can discover the desired optimal conditions with fewer trials than ever before.

Epistra-Acc-3_edited.png

*Experiment automation equipment: It is possible to support various equipment such as automatic pipetting machines (robots), parallel microbioreactors (Ambr, etc.), and process control at manufacturing plants.

Epistra-AI-a.png

Automatic condition assessment AI

With AI that utilizes deep learning technology, it is possible to reproduce cell state judgments that are at least as good as those of experts. In this example, about 100 microscopic images were used as training data, and we were able to distinguish between good and bad cells with an accuracy of over 80%.

Automated experiment planning AI

AI enhanced to solve life science problems, implementing our own algorithm based on Bayesian optimization technology. There are three problems (high dimensionality, high noise, and high cost) in applying standard Bayesian optimization technology to life science problems, and we solve them with our own algorithm.

Epistra-AI-b.png

CASE STUDY

IMG_0565-2.jpg

The superiority of our technology is being proven by multiple cases. Please contact us for details.

Human iPS cells

Optimization of RPE differentiation induction protocol

[RIKEN Center for Biosystems Dynamics Research (BDR)]

Schedule: from April 2018

Goal: Achieving a differentiation efficiency that surpasses masters with robots and AI

Status: paper published (news release)

Microbial culture process optimization

[Biotech company]

Schedule: January to June 2020

Goal: Discover culture conditions that maximize the yield of the desired product

Status: PoC completed (news release)

Automatic determination of findings from medical image information

[Kobe eye center hospital]

Schedule: February to April 2019

Goal: PoC for automating finding judgments used to assist doctors in diagnosing

status:Paper published (news release)

bottom of page