DiffER: Categorical Diffusion for Chemical Retrosynthesis

TL;DR

DiffER introduces a novel template-free categorical diffusion approach for chemical retrosynthesis, predicting entire SMILES sequences simultaneously and achieving state-of-the-art accuracy among non-template methods.

Contribution

The paper presents DiffER, a new diffusion-based model for retrosynthesis that outperforms existing template-free methods and includes a length prediction component for improved accuracy.

Findings

DiffER achieves state-of-the-art top-1 accuracy among template-free methods.

The ensemble of diffusion models improves confidence and likelihood in predictions.

Accurate SMILES length prediction is crucial for model performance.

Abstract

Methods for automatic chemical retrosynthesis have found recent success through the application of models traditionally built for natural language processing, primarily through transformer neural networks. These models have demonstrated significant ability to translate between the SMILES encodings of chemical products and reactants, but are constrained as a result of their autoregressive nature. We propose DiffER, an alternative template-free method for retrosynthesis prediction in the form of categorical diffusion, which allows the entire output SMILES sequence to be predicted in unison. We construct an ensemble of diffusion models which achieves state-of-the-art performance for top-1 accuracy and competitive performance for top-3, top-5, and top-10 accuracy among template-free methods. We prove that DiffER is a strong baseline for a new class of template-free model, capable of learning a variety of synthetic techniques used in laboratory settings and outperforming a variety of other template-free methods on top-k accuracy metrics. By constructing an ensemble of categorical diffusion models with a novel length prediction component with variance, our method is able to approximately sample from the posterior distribution of reactants, producing results with strong metrics of confidence and likelihood. Furthermore, our analyses demonstrate that accurate prediction of the SMILES sequence length is key to further boosting the performance of categorical diffusion models.