GeneDisco: A Benchmark for Experimental Design in Drug Discovery

Published in ICLR, 2022

Recommended citation: Mehrjou A, Soleymani A, Jesson A, Notin P, Gal Y, Bauer S, Schwab P. GeneDisco: A Benchmark for Experimental Design in Drug Discovery. arXiv preprint arXiv:2110.11875. 2021 Oct 22. https://tinyurl.com/3nec97kb

We introduce GeneDisco, an open benchmark suite for batch active learning for drug discovery that provides curated datasets, tasks, performance evaluation and open source implementations of state-of-the-art algorithms for experimental exploration.

  • We perform an extensive experimental baseline evaluation that establishes the relative performance of existing state-of-the-art methods on all the developed benchmark settings using a total of more than 20 000 central processing unit (CPU) hours of compute time.
  • We survey and analyse the current state-of-the-art of active learning for biological exploration in the context of the generated experimental results, and present avenues of heightened potential for future research based on the developed benchmark.

drawing

Download paper here

Leave a Comment