English

Bootstrapped Training of Score-Conditioned Generator for Offline Design of Biological Sequences

Quantitative Methods 2024-03-26 v2 Machine Learning Machine Learning

Abstract

We study the problem of optimizing biological sequences, e.g., proteins, DNA, and RNA, to maximize a black-box score function that is only evaluated in an offline dataset. We propose a novel solution, bootstrapped training of score-conditioned generator (BootGen) algorithm. Our algorithm repeats a two-stage process. In the first stage, our algorithm trains the biological sequence generator with rank-based weights to enhance the accuracy of sequence generation based on high scores. The subsequent stage involves bootstrapping, which augments the training dataset with self-generated data labeled by a proxy score function. Our key idea is to align the score-based generation with a proxy score function, which distills the knowledge of the proxy score function to the generator. After training, we aggregate samples from multiple bootstrapped generators and proxies to produce a diverse design. Extensive experiments show that our method outperforms competitive baselines on biological sequential design tasks. We provide reproducible source code: \href{https://github.com/kaist-silab/bootgen}{https://github.com/kaist-silab/bootgen}.

Keywords

Cite

@article{arxiv.2306.03111,
  title  = {Bootstrapped Training of Score-Conditioned Generator for Offline Design of Biological Sequences},
  author = {Minsu Kim and Federico Berto and Sungsoo Ahn and Jinkyoo Park},
  journal= {arXiv preprint arXiv:2306.03111},
  year   = {2024}
}

Comments

NeurIPS 2023, 19 pages, 5 figures