De novo Drug Design using Reinforcement Learning with Multiple GPT Agents

Xiuyuan Hu, Guoqing Liu, Yang Zhao, Hao Zhang
2/12/2026

No official code found. Best implementation: HXYfighter/MolRL-MGPT (40 stars, last commit 2024-03-27).

Abstract

De novo drug design is a pivotal issue in pharmacology and a new area of focus in AI for science research. A central challenge in this field is to generate molecules with specific properties while also producing a wide range of diverse candidates. Although advanced technologies such as transformer models and reinforcement learning have been applied in drug design, their potential has not been fully realized. Therefore, we propose MolRL-MGPT, a reinforcement learning algorithm with multiple GPT agents for drug molecular generation. To promote molecular diversity, we encourage the agents to collaborate in searching for desirable molecules in diverse directions. Our algorithm has shown promising results on the GuacaMol benchmark and exhibits efficacy in designing inhibitors against SARS-CoV-2 protein targets. The codes are available at: https://github.com/HXYfighter/MolRL-MGPT.

View on arXivView PDF

Code Implementations(1)

HXYfighter/MolRL-MGPTCommunity100%

NeurIPS 2023 paper: De novo Drug Design using Reinforcement Learning with Multiple GPT Agents

407PythonOct 7, 20232 years ago

Similar Papers

Utilizing Reinforcement Learning for de novo Drug Design

88% match2023

Improving Targeted Molecule Generation through Language Model Fine-Tuning Via Reinforcement Learning

86% match2024

ReACT-Drug: Reaction-Template Guided Reinforcement Learning for de novo Drug Design

84% match1 repo42025

A Survey of Generative AI for de novo Drug Design: New Frontiers in Molecule and Protein Generation

83% match1 repo952024

Mol-AIR: Molecular Reinforcement Learning with Adaptive Intrinsic Rewards for Goal-directed Molecular Generation

83% match1 repo292024

Active Learning-Guided Seq2Seq Variational Autoencoder for Multi-target Inhibitor Generation

82% match2025

Cite this paper

@article{hu2026novo,
  title  = {De novo Drug Design using Reinforcement Learning with Multiple GPT Agents},
  author = {Xiuyuan Hu and Guoqing Liu and Yang Zhao and Hao Zhang},
  year   = {2026},
  eprint = {2401.06155},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/2401.06155},
  journal = {NEURIPS 2023 2023}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2401.06155.svg)](https://codeofpaper.com/paper/2401.06155)

Discussion