Saved in:
Bibliographic Details
Main Authors: Yang, Qianwei, Xu, Dong, Yang, Zhangfan, Yuan, Sisi, Zhu, Zexuan, Li, Jianqiang, Ji, Junkai
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2601.21964
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918314198958080
author Yang, Qianwei
Xu, Dong
Yang, Zhangfan
Yuan, Sisi
Zhu, Zexuan
Li, Jianqiang
Ji, Junkai
author_facet Yang, Qianwei
Xu, Dong
Yang, Zhangfan
Yuan, Sisi
Zhu, Zexuan
Li, Jianqiang
Ji, Junkai
contents Drug discovery can be viewed as a combinatorial search over an immense chemical space, motivating the development of deep generative models for de novo molecular design. Among these, GPT-based molecular language models (MLM) have shown strong molecular design performance by learning chemical syntax and semantics from large-scale data. However, existing MLMs face two fundamental limitations: they inadequately capture the graph-structured nature of molecules when formulated as next-token prediction problems, and they typically lack explicit mechanisms for target-aware generation. Here, we propose SoftMol, a unified framework that co-designs molecular representation, model architecture, and search strategy for target-aware molecular generation. SoftMol introduces soft fragments, a rule-free block representation of SMILES that enables diffusion-native modeling, and develops SoftBD, the first block-diffusion molecular language model that combines local bidirectional diffusion with autoregressive generation under molecular structural constraints. To favor generated molecules with high drug-likeness and synthetic accessibility, SoftBD is trained on a carefully curated dataset named ZINC-Curated. SoftMol further integrates a gated Monte Carlo tree search to assemble fragments in a target-aware manner. Experimental results show that, compared with current state-of-the-art models, SoftMol achieves 100% chemical validity, improves binding affinity by 9.7%, yields a 2-3x increase in molecular diversity, and delivers a 6.6x speedup in inference efficiency. Code is available at https://github.com/szu-aicourse/softmol
format Preprint
id arxiv_https___arxiv_org_abs_2601_21964
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle From Tokens to Blocks: A Block-Diffusion Perspective on Molecular Generation
Yang, Qianwei
Xu, Dong
Yang, Zhangfan
Yuan, Sisi
Zhu, Zexuan
Li, Jianqiang
Ji, Junkai
Machine Learning
Drug discovery can be viewed as a combinatorial search over an immense chemical space, motivating the development of deep generative models for de novo molecular design. Among these, GPT-based molecular language models (MLM) have shown strong molecular design performance by learning chemical syntax and semantics from large-scale data. However, existing MLMs face two fundamental limitations: they inadequately capture the graph-structured nature of molecules when formulated as next-token prediction problems, and they typically lack explicit mechanisms for target-aware generation. Here, we propose SoftMol, a unified framework that co-designs molecular representation, model architecture, and search strategy for target-aware molecular generation. SoftMol introduces soft fragments, a rule-free block representation of SMILES that enables diffusion-native modeling, and develops SoftBD, the first block-diffusion molecular language model that combines local bidirectional diffusion with autoregressive generation under molecular structural constraints. To favor generated molecules with high drug-likeness and synthetic accessibility, SoftBD is trained on a carefully curated dataset named ZINC-Curated. SoftMol further integrates a gated Monte Carlo tree search to assemble fragments in a target-aware manner. Experimental results show that, compared with current state-of-the-art models, SoftMol achieves 100% chemical validity, improves binding affinity by 9.7%, yields a 2-3x increase in molecular diversity, and delivers a 6.6x speedup in inference efficiency. Code is available at https://github.com/szu-aicourse/softmol
title From Tokens to Blocks: A Block-Diffusion Perspective on Molecular Generation
topic Machine Learning
url https://arxiv.org/abs/2601.21964