Swallowing the Bitter Pill: Simplified Scalable Conformer Generation

Yuyang Wang, Ahmed A. A. Elhag, Navdeep Jaitly, Joshua M. Susskind, Miguel Ángel Bautista
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:50400-50418, 2024.

Abstract

We present a novel way to predict molecular conformers through a simple formulation that sidesteps many of the heuristics of prior works and achieves state of the art results by using the advantages of scale. By training a diffusion generative model directly on 3D atomic positions without making assumptions about the explicit structure of molecules (e.g. modeling torsional angles) we are able to radically simplify structure learning, and make it trivial to scale up the model sizes. This model, called Molecular Conformer Fields (MCF), works by parameterizing conformer structures as functions that map elements from a molecular graph directly to their 3D location in space. This formulation allows us to boil down the essence of structure prediction to learning a distribution over functions. Experimental results show that scaling up the model capacity leads to large gains in generalization performance without enforcing inductive biases like rotational equivariance. MCF represents an advance in extending diffusion models to handle complex scientific problems in a conceptually simple, scalable and effective manner.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-wang24q, title = {Swallowing the Bitter Pill: Simplified Scalable Conformer Generation}, author = {Wang, Yuyang and Elhag, Ahmed A. A. and Jaitly, Navdeep and Susskind, Joshua M. and Bautista, Miguel \'{A}ngel}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {50400--50418}, year = {2024}, editor = {Salakhutdinov, Ruslan and Kolter, Zico and Heller, Katherine and Weller, Adrian and Oliver, Nuria and Scarlett, Jonathan and Berkenkamp, Felix}, volume = {235}, series = {Proceedings of Machine Learning Research}, month = {21--27 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v235/main/assets/wang24q/wang24q.pdf}, url = {https://proceedings.mlr.press/v235/wang24q.html}, abstract = {We present a novel way to predict molecular conformers through a simple formulation that sidesteps many of the heuristics of prior works and achieves state of the art results by using the advantages of scale. By training a diffusion generative model directly on 3D atomic positions without making assumptions about the explicit structure of molecules (e.g. modeling torsional angles) we are able to radically simplify structure learning, and make it trivial to scale up the model sizes. This model, called Molecular Conformer Fields (MCF), works by parameterizing conformer structures as functions that map elements from a molecular graph directly to their 3D location in space. This formulation allows us to boil down the essence of structure prediction to learning a distribution over functions. Experimental results show that scaling up the model capacity leads to large gains in generalization performance without enforcing inductive biases like rotational equivariance. MCF represents an advance in extending diffusion models to handle complex scientific problems in a conceptually simple, scalable and effective manner.} }
Endnote
%0 Conference Paper %T Swallowing the Bitter Pill: Simplified Scalable Conformer Generation %A Yuyang Wang %A Ahmed A. A. Elhag %A Navdeep Jaitly %A Joshua M. Susskind %A Miguel Ángel Bautista %B Proceedings of the 41st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Ruslan Salakhutdinov %E Zico Kolter %E Katherine Heller %E Adrian Weller %E Nuria Oliver %E Jonathan Scarlett %E Felix Berkenkamp %F pmlr-v235-wang24q %I PMLR %P 50400--50418 %U https://proceedings.mlr.press/v235/wang24q.html %V 235 %X We present a novel way to predict molecular conformers through a simple formulation that sidesteps many of the heuristics of prior works and achieves state of the art results by using the advantages of scale. By training a diffusion generative model directly on 3D atomic positions without making assumptions about the explicit structure of molecules (e.g. modeling torsional angles) we are able to radically simplify structure learning, and make it trivial to scale up the model sizes. This model, called Molecular Conformer Fields (MCF), works by parameterizing conformer structures as functions that map elements from a molecular graph directly to their 3D location in space. This formulation allows us to boil down the essence of structure prediction to learning a distribution over functions. Experimental results show that scaling up the model capacity leads to large gains in generalization performance without enforcing inductive biases like rotational equivariance. MCF represents an advance in extending diffusion models to handle complex scientific problems in a conceptually simple, scalable and effective manner.
APA
Wang, Y., Elhag, A.A.A., Jaitly, N., Susskind, J.M. & Bautista, M.Á.. (2024). Swallowing the Bitter Pill: Simplified Scalable Conformer Generation. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:50400-50418 Available from https://proceedings.mlr.press/v235/wang24q.html.

Related Material