Enregistré dans:
Détails bibliographiques
Auteurs principaux: Thomas, Morgan, Bou, Albert, De Fabritiis, Gianni
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:https://arxiv.org/abs/2501.19153
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866917120432930816
author Thomas, Morgan
Bou, Albert
De Fabritiis, Gianni
author_facet Thomas, Morgan
Bou, Albert
De Fabritiis, Gianni
contents Chemical Language Models (CLMs) leveraging reinforcement learning (RL) have shown promise in de novo molecular design, yet often suffer from mode collapse, limiting their exploration capabilities. Inspired by Test-Time Training (TTT) in large language models, we propose scaling TTT for CLMs to enhance chemical space exploration. We introduce MolExp, a novel benchmark emphasizing the discovery of structurally diverse molecules with similar bioactivity, simulating real-world drug design challenges. Our results demonstrate that scaling TTT by increasing the number of independent RL agents follows a log-linear scaling law, significantly improving exploration efficiency as measured by MolExp. In contrast, increasing TTT training time yields diminishing returns, even with exploration bonuses. We further evaluate cooperative RL strategies to enhance exploration efficiency. These findings provide a scalable framework for generative molecular design, offering insights into optimizing AI-driven drug discovery.
format Preprint
id arxiv_https___arxiv_org_abs_2501_19153
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Test-Time Training Scaling Laws for Chemical Exploration in Drug Design
Thomas, Morgan
Bou, Albert
De Fabritiis, Gianni
Machine Learning
Chemical Language Models (CLMs) leveraging reinforcement learning (RL) have shown promise in de novo molecular design, yet often suffer from mode collapse, limiting their exploration capabilities. Inspired by Test-Time Training (TTT) in large language models, we propose scaling TTT for CLMs to enhance chemical space exploration. We introduce MolExp, a novel benchmark emphasizing the discovery of structurally diverse molecules with similar bioactivity, simulating real-world drug design challenges. Our results demonstrate that scaling TTT by increasing the number of independent RL agents follows a log-linear scaling law, significantly improving exploration efficiency as measured by MolExp. In contrast, increasing TTT training time yields diminishing returns, even with exploration bonuses. We further evaluate cooperative RL strategies to enhance exploration efficiency. These findings provide a scalable framework for generative molecular design, offering insights into optimizing AI-driven drug discovery.
title Test-Time Training Scaling Laws for Chemical Exploration in Drug Design
topic Machine Learning
url https://arxiv.org/abs/2501.19153