Alignment is Key for Applying Diffusion Models to Retrosynthesis
- URL: http://arxiv.org/abs/2405.17656v1
- Date: Mon, 27 May 2024 20:57:19 GMT
- Title: Alignment is Key for Applying Diffusion Models to Retrosynthesis
- Authors: Najwa Laabid, Severi Rissanen, Markus Heinonen, Arno Solin, Vikas Garg,
- Abstract summary: Diffusion models are a promising modelling approach, enabling post-hoc conditioning and trading off quality for speed during generation.
We show mathematically that permutation equivariant denoisers severely limit the expressiveness of graph diffusion models and thus their adaptation to retrosynthesis.
Our new denoiser achieves the highest top-$1$ accuracy ($54.7$%) across template-free and template-based methods on USPTO-50k.
- Score: 24.912841472542322
- License: http://creativecommons.org/licenses/by/4.0/
- Abstract: Retrosynthesis, the task of identifying precursors for a given molecule, can be naturally framed as a conditional graph generation task. Diffusion models are a particularly promising modelling approach, enabling post-hoc conditioning and trading off quality for speed during generation. We show mathematically that permutation equivariant denoisers severely limit the expressiveness of graph diffusion models and thus their adaptation to retrosynthesis. To address this limitation, we relax the equivariance requirement such that it only applies to aligned permutations of the conditioning and the generated graphs obtained through atom mapping. Our new denoiser achieves the highest top-$1$ accuracy ($54.7$\%) across template-free and template-based methods on USPTO-50k. We also demonstrate the ability for flexible post-training conditioning and good sample quality with small diffusion step counts, highlighting the potential for interactive applications and additional controls for multi-step planning.
Related papers
- BindGPT: A Scalable Framework for 3D Molecular Design via Language Modeling and Reinforcement Learning [11.862370962277938]
We present a novel generative model, BindGPT, which uses a conceptually simple but powerful approach to create 3D molecules within the protein's binding site.
We show how such simple conceptual approach combined with pretraining and scaling can perform on par or better than the current best specialized diffusion models.
arXiv Detail & Related papers (2024-06-06T02:10:50Z) - Graph Diffusion Transformer for Multi-Conditional Molecular Generation [16.58392955245203]
We present the Graph Diffusion Transformer (Graph DiT) for multi-conditional molecular generation.
Graph DiT has a condition encoder to learn the representation of numerical and categorical properties.
We extensively validate the Graph DiT for multi-conditional polymer and small molecule generation.
arXiv Detail & Related papers (2024-01-24T23:45:31Z) - Towards Controllable Diffusion Models via Reward-Guided Exploration [15.857464051475294]
We propose a novel framework that guides the training-phase of diffusion models via reinforcement learning (RL)
RL enables calculating policy gradients via samples from a pay-off distribution proportional to exponential scaled rewards, rather than from policies themselves.
Experiments on 3D shape and molecule generation tasks show significant improvements over existing conditional diffusion models.
arXiv Detail & Related papers (2023-04-14T13:51:26Z) - ChiroDiff: Modelling chirographic data with Diffusion Models [132.5223191478268]
We introduce a powerful model-class namely "Denoising Diffusion Probabilistic Models" or DDPMs for chirographic data.
Our model named "ChiroDiff", being non-autoregressive, learns to capture holistic concepts and therefore remains resilient to higher temporal sampling rate.
arXiv Detail & Related papers (2023-04-07T15:17:48Z) - ShiftDDPMs: Exploring Conditional Diffusion Models by Shifting Diffusion
Trajectories [144.03939123870416]
We propose a novel conditional diffusion model by introducing conditions into the forward process.
We use extra latent space to allocate an exclusive diffusion trajectory for each condition based on some shifting rules.
We formulate our method, which we call textbfShiftDDPMs, and provide a unified point of view on existing related methods.
arXiv Detail & Related papers (2023-02-05T12:48:21Z) - Modular Flows: Differential Molecular Generation [18.41106104201439]
Flows can generate molecules effectively by inverting the encoding process.
Existing flow models require artifactual dequantization or specific node/edge orderings.
We develop continuous normalizing E(3)-equivariant flows, based on a system of node ODEs and a graph PDE.
Our models can be cast as message-passing temporal networks, and result in superlative performance on the tasks of density estimation and molecular generation.
arXiv Detail & Related papers (2022-10-12T09:08:35Z) - DiGress: Discrete Denoising diffusion for graph generation [79.13904438217592]
DiGress is a discrete denoising diffusion model for generating graphs with categorical node and edge attributes.
It achieves state-of-the-art performance on molecular and non-molecular datasets, with up to 3x validity improvement.
It is also the first model to scale to the large GuacaMol dataset containing 1.3M drug-like molecules.
arXiv Detail & Related papers (2022-09-29T12:55:03Z) - GeoDiff: a Geometric Diffusion Model for Molecular Conformation
Generation [102.85440102147267]
We propose a novel generative model named GeoDiff for molecular conformation prediction.
We show that GeoDiff is superior or comparable to existing state-of-the-art approaches.
arXiv Detail & Related papers (2022-03-06T09:47:01Z) - Score-based Generative Modeling of Graphs via the System of Stochastic
Differential Equations [57.15855198512551]
We propose a novel score-based generative model for graphs with a continuous-time framework.
We show that our method is able to generate molecules that lie close to the training distribution yet do not violate the chemical valency rule.
arXiv Detail & Related papers (2022-02-05T08:21:04Z) - Denoising Diffusion Probabilistic Models [91.94962645056896]
We present high quality image synthesis results using diffusion probabilistic models.
Our best results are obtained by training on a weighted variational bound designed according to a novel connection between diffusion probabilistic models and denoising score matching with Langevin dynamics.
arXiv Detail & Related papers (2020-06-19T17:24:44Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.