AI & Machine Learning

We provide data, insights, and technologies needed to accelerate drug discovery, drive materials innovation and find better starting places.

We provide data, insights, and technologies needed to accelerate drug discovery, drive materials innovation and find better starting places.

CONTACT US

Improve prediction accuracy

Data, descriptors, and algorithms all play critical roles in machine learning and must be carefully optimized to maximize prediction accuracy.

Discover new chemistry

Models cannot predict what they haven’t seen. Data diversity and representation, even in sparsely populated subspaces, enable more novel predictions.

Find more viable starting places

Identify more promising leads and targets to reduce the need for unnecessary trials and experiments required to improve R&D efficiency.

Molecular representation is the most important part of AI. If you get it wrong, you are lost from the beginning. It's even more important than the algorithm.

Jürgen Cox, Ph.D.
Professor
Max Planck Institute of Biochemistry
Path 170 Copy 4

Predicting New Chemistry

Bayer scientists increased prediction accuracy by 32 percentage points with scientist-curated data from CAS. Enhanced predictive power in rare reaction classes contributes new, useful results to open up difficult areas of science.

Download Whitepaper
Design Element 4

ICL uncovers new uses for waste compounds

“It was imperative that AI could efficiently link our target compounds to known applications in the data.” 

Aviad Karniel; Operations manager
Central R&D Facility, ICL
Read the ICL Case Study
Design Element 4

Impact of Data Quality on Machine Learning Results

See how enhancing the quality of the input data alone, without changing the algorithm, impacts prediction accuracy of an algorithm designed to assess the biological activity of compounds relative to different targets in this quantitative study by CAS data scientists. 

Download the Case Study
Design Element 4

"More and better data is obvious.

What isn’t obvious is how data is interpreted. Encoding requires expert chemistry knowledge. CAS descriptors get better results because of the expertise of their scientists".

Dr. Alpha Lee, Co-founder/Chief Scientific Officer
PostEra
Design Element 4

Our Services

Leverage our scientific expertise, comprehensive content collection, and highly structured, big data environment to power the success of your AI and machine learning intiatives.

Custom curation

Meet your specific model requirements with hard-to-find data curated by our scientists

Harmonization and structuring

Ensure your data is AI-ready

Content licensing

Train and validate your model with the CAS Content Collection

Customized molecular descriptors

Optimally encode and represent chemical structures

Custom consulting

Avoid common missteps with scientific information for informatics success

Resources

More resources