Coarse-to-Fine: a Hierarchical Diffusion Model for Molecule Generation in 3D

Bo Qiang, Yuxuan Song, Minkai Xu, Jingjing Gong, Bowen Gao, Hao Zhou, Wei-Ying Ma, Yanyan Lan
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:28277-28299, 2023.

Abstract

Generating desirable molecular structures in 3D is a fundamental problem for drug discovery. Despite the considerable progress we have achieved, existing methods usually generate molecules in atom resolution and ignore intrinsic local structures such as rings, which leads to poor quality in generated structures, especially when generating large molecules. Fragment-based molecule generation is a promising strategy, however, it is nontrivial to be adapted for 3D non-autoregressive generations because of the combinational optimization problems. In this paper, we utilize a coarse-to-fine strategy to tackle this problem, in which a Hierarchical Diffusion-based model (i.e. HierDiff) is proposed to preserve the validity of local segments without relying on autoregressive modeling. Specifically, HierDiff first generates coarse-grained molecule geometries via an equivariant diffusion process, where each coarse-grained node reflects a fragment in a molecule. Then the coarse-grained nodes are decoded into fine-grained fragments by a message-passing process and a newly designed iterative refined sampling module. Lastly, the fine-grained fragments are then assembled to derive a complete atomic molecular structure. Extensive experiments demonstrate that HierDiff consistently improves the quality of molecule generation over existing methods.

Cite this Paper

BibTeX
@InProceedings{pmlr-v202-qiang23a, title = {Coarse-to-Fine: a Hierarchical Diffusion Model for Molecule Generation in 3{D}}, author = {Qiang, Bo and Song, Yuxuan and Xu, Minkai and Gong, Jingjing and Gao, Bowen and Zhou, Hao and Ma, Wei-Ying and Lan, Yanyan}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {28277--28299}, year = {2023}, editor = {Krause, Andreas and Brunskill, Emma and Cho, Kyunghyun and Engelhardt, Barbara and Sabato, Sivan and Scarlett, Jonathan}, volume = {202}, series = {Proceedings of Machine Learning Research}, month = {23--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v202/qiang23a/qiang23a.pdf}, url = {https://proceedings.mlr.press/v202/qiang23a.html}, abstract = {Generating desirable molecular structures in 3D is a fundamental problem for drug discovery. Despite the considerable progress we have achieved, existing methods usually generate molecules in atom resolution and ignore intrinsic local structures such as rings, which leads to poor quality in generated structures, especially when generating large molecules. Fragment-based molecule generation is a promising strategy, however, it is nontrivial to be adapted for 3D non-autoregressive generations because of the combinational optimization problems. In this paper, we utilize a coarse-to-fine strategy to tackle this problem, in which a Hierarchical Diffusion-based model (i.e. HierDiff) is proposed to preserve the validity of local segments without relying on autoregressive modeling. Specifically, HierDiff first generates coarse-grained molecule geometries via an equivariant diffusion process, where each coarse-grained node reflects a fragment in a molecule. Then the coarse-grained nodes are decoded into fine-grained fragments by a message-passing process and a newly designed iterative refined sampling module. Lastly, the fine-grained fragments are then assembled to derive a complete atomic molecular structure. Extensive experiments demonstrate that HierDiff consistently improves the quality of molecule generation over existing methods.} }
Endnote
%0 Conference Paper %T Coarse-to-Fine: a Hierarchical Diffusion Model for Molecule Generation in 3D %A Bo Qiang %A Yuxuan Song %A Minkai Xu %A Jingjing Gong %A Bowen Gao %A Hao Zhou %A Wei-Ying Ma %A Yanyan Lan %B Proceedings of the 40th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2023 %E Andreas Krause %E Emma Brunskill %E Kyunghyun Cho %E Barbara Engelhardt %E Sivan Sabato %E Jonathan Scarlett %F pmlr-v202-qiang23a %I PMLR %P 28277--28299 %U https://proceedings.mlr.press/v202/qiang23a.html %V 202 %X Generating desirable molecular structures in 3D is a fundamental problem for drug discovery. Despite the considerable progress we have achieved, existing methods usually generate molecules in atom resolution and ignore intrinsic local structures such as rings, which leads to poor quality in generated structures, especially when generating large molecules. Fragment-based molecule generation is a promising strategy, however, it is nontrivial to be adapted for 3D non-autoregressive generations because of the combinational optimization problems. In this paper, we utilize a coarse-to-fine strategy to tackle this problem, in which a Hierarchical Diffusion-based model (i.e. HierDiff) is proposed to preserve the validity of local segments without relying on autoregressive modeling. Specifically, HierDiff first generates coarse-grained molecule geometries via an equivariant diffusion process, where each coarse-grained node reflects a fragment in a molecule. Then the coarse-grained nodes are decoded into fine-grained fragments by a message-passing process and a newly designed iterative refined sampling module. Lastly, the fine-grained fragments are then assembled to derive a complete atomic molecular structure. Extensive experiments demonstrate that HierDiff consistently improves the quality of molecule generation over existing methods.
APA
Qiang, B., Song, Y., Xu, M., Gong, J., Gao, B., Zhou, H., Ma, W. & Lan, Y.. (2023). Coarse-to-Fine: a Hierarchical Diffusion Model for Molecule Generation in 3D. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:28277-28299 Available from https://proceedings.mlr.press/v202/qiang23a.html.

Related Material