#machine learning #computer vision #climate science #quantum computing #human-computer interaction #data science #generative ai #reinforcement learning #federated learning #edge computing #gene editing

SynGFN: learning across chemical space with generative flow-based molecular discovery

Yuchen Zhu Shuwang Li Jihong Chen Donghai Zhao Xiaorui Wang Yitong Li Yifei Liu Yue Kong Beichen Zhang Chang Liu Tingjun Hou Chang-Yu Hsieh

Nov 2025 · Nature Computational Science · Vol 6, pp. 29 - 38 · 3 citations · 49 references

Computer Science Medicine

TL;DR

This study proposes SynGFN, which models molecular design as a cascade of simulated chemical reactions, enabling the assembly of molecules from synthesizable building blocks, as a bridge linking molecular design and synthesis, accelerating exploration and producing diverse, synthesizable, high-performance molecules.

DOI PubMed Semantic Scholar

View source

Similar papers

Book Open access Aug 2026

From VAEs to Diffusion and LLMs: Modern Generative Models for Molecular Discovery

This tutorial provides a unified introduction to modern generative modeling approaches for molecular generation, covering their theoretical foundations, algorithmic design, and practical considerations for molecular representations such as 1D SMILES strings, 2D molecular graphs, and 3D structures.

Ke-Han Guo, Yili Shen, Jeehyun Hwang et al. · Proceedings of the 32nd ACM... · 0 citations

Review Aug 2026

The Evolution of Generative Chemistry in Medicinal Chemistry

A critical perspective is provided on how generative models are shaping the future of rational and reliable drug design, including automated synthesis planning, retrosynthesis prediction, and multi‐objective optimization.

Rania Ehab Koshty, Manar Ahmed Shehata, Ahmed M. Gab Allah et al. · ChemistrySelect · 0 citations

Jul 2026

A unified deep generative framework for surrogate-guided molecular discovery across diverse molecular spaces.

A high-accuracy surrogate predictor based on message-passing neural networks is decoupled from heterogeneous generators, and integrated into a closed-loop strategy of generate, score, select, and regenerate to progressively enrich high-performing candidates without modifying the underlying model architectures.

Zhao-Sheng Zhang, Yan-Bo Liu, Jia-Dong Liu et al. · Physical Chemistry, Chemical... · 0 citations

Open access Aug 2026

Generalizing molecular design via flexible search space control

ChemTSv3 is introduced, an exploration framework with a flexible architecture that accommodates diverse design scenarios for adaptive molecular design, and shows that this flexibility enables efficient exploration across diverse design spaces, from drug-like small molecules to proteins.

Satoru Fujii, Yuki Murakami, Tatsuya Yoshizawa et al. · Communications Chemistry · 2 citations

#artificial intelligence Preprint Aug 2026

PGFS++: Molecular Property Improvement under Synthesis and Diversity Constraints

PGFS++ is introduced, a synthesis-aware reinforcement learning framework for input-specific molecular improvement that improves target properties while preserving high output diversity, and experiments show that PGFS++ improves target properties while preserving high output diversity.

Bo-Qi Zhang, G. James, S. Gottipati et al. · 0 citations

#machine learning Preprint Sep 2026

Training Large Language Models for Small-Molecule Design with Synthetic Task Scaling

This work investigates whether LLMs can learn molecular design strategies from cheaper synthetic tasks that generalize to expensive molecular lead optimization settings and suggests that scaling post-training using synthetic tasks is an effective strategy for adapting LLMs to high-cost experimental scenarios that are t...

Frank Hu, Shriram Chennakesavalu, Zi-Cheng Wang et al. · 0 citations

SynGFN: learning across chemical space with generative flow-based molecular discovery

Similar papers

From VAEs to Diffusion and LLMs: Modern Generative Models for Molecular Discovery

The Evolution of Generative Chemistry in Medicinal Chemistry

A unified deep generative framework for surrogate-guided molecular discovery across diverse molecular spaces.

Generalizing molecular design via flexible search space control

PGFS++: Molecular Property Improvement under Synthesis and Diversity Constraints

Training Large Language Models for Small-Molecule Design with Synthetic Task Scaling

Related blog posts

What AI gets wrong and what failure teaches us

Alexander Rakhlin named director of the MIT Statistics and Data Science Center

Towards a quantum computer that learns from its errors

MIT Schwarzman College of Computing launches pilot to help educators teach AI across disciplines