Improving Few- and Zero-Shot Reaction Template Prediction Using Modern Hopfield Networks (2022-01-15T00:00:00.000000Z)

TL;DR

A template-based single-step retrosynthesis model based on Modern Hopfield Networks is introduced, which learn an encoding of both molecules and reaction templates in order to predict the relevance of templates for a given molecule.

Abstract

Finding synthesis routes for molecules of interest is essential in the discovery of new drugs and materials. To find such routes, computer-assisted synthesis planning (CASP) methods are employed, which rely on a single-step model of chemical reactivity. In this study, we introduce a template-based single-step retrosynthesis model based on Modern Hopfield Networks, which learn an encoding of both molecules and reaction templates in order to predict the relevance of templates for a given molecule. The template representation allows generalization across different reactions and significantly improves the performance of template relevance prediction, especially for templates with few or zero training examples. With inference speed up to orders of magnitude faster than baseline methods, we improve or match the state-of-the-art performance for top-k exact match accuracy for k ≥ 3 in the retrosynthesis benchmark USPTO-50k. Code to reproduce the results is available at github.com/ml-jku/mhn-react.

Authors

Sepp Hochreiter

12 papers

Marwin H. S. Segler

7 papers

G. Klambauer

7 papers

TL;DR

Abstract

Authors

References87 items

RetCL: A Selection-based Approach for Retrosynthesis via Contrastive Learning

RetroPrime: A Diverse, Plausible and Transformer-based Method for Single-Step Retrosynthesis Predictions

Retrosynthesis prediction using grammar-based neural machine translation: An information-theoretic approach

Accuracy-diversity trade-off in recommender systems via graph convolutions

Learning Transferable Visual Models From Natural Language Supervision

Single-Step Retrosynthesis Prediction Based on the Identification of Potential Disconnection Sites Using Molecular Substructure Fingerprints

Barking up the right tree: an approach to search over molecule synthesis DAGs

Interpretable machine-learning strategy for soft-magnetic property and thermal stability in Fe-based metallic glasses

Contrastive Learning of Medical Visual Representations from Paired Images and Text

Data Transfer Approaches to Improve Seq-to-Seq Retrosynthesis

Substructure-based neural machine translation for retrosynthetic prediction

Machine learning the ropes: principles, applications and directions in synthetic chemistry.

Retro*: Learning Retrosynthetic Planning with Neural Guided A* Search

Molecule Edit Graph Attention Network: Modeling Chemical Reactions as Sequences of Graph Edits

RetroXpert: Decompose Retrosynthesis Prediction like a Chemist

Learning Graph Models for Template-Free Retrosynthesis

Artificial applicability labels for improving policies in retrosynthesis prediction

Molecular Design in Synthetically Accessible Chemical Space via Deep Reinforcement Learning

Learning To Navigate The Synthetically Accessible Chemical Space Using Reinforcement Learning

Modern Hopfield Networks and Attention for Immune Repertoire Classification

Current and Future Roles of Artificial Intelligence in Medicinal Chemistry Synthesis

A Graph to Graphs Framework for Retrosynthesis Prediction

A Bayesian algorithm for retrosynthesis

State-of-the-art augmented NLP transformer models for direct and single-step retrosynthesis

A Simple Framework for Contrastive Learning of Visual Representations

Data Augmentation and Pretraining for Template-Based Retrosynthetic Prediction in Computer-Aided Synthesis Planning

Retrosynthesis Prediction with Conditional Graph Logic Network

World Programs for Model-Based Learning and Planning in Compositional State and Action Spaces

Predicting Retrosynthetic Reactions Using Self-Corrected Transformer Neural Networks

Self-Supervised Learning of Pretext-Invariant Representations

PyTorch: An Imperative Style, High-Performance Deep Learning Library

On failure modes in molecule generation and optimization.

Momentum Contrast for Unsupervised Visual Representation Learning

Molecular Graph Enhanced Transformer for Retrosynthesis Prediction

Learning to Make Generalizable and Diverse Predictions for Retrosynthesis

Prediction and Interpretable Visualization of Retrosynthetic Reactions Using Graph Convolutional Networks

Contrastive Multiview Coding

A Model to Search for Synthesizable Molecules

A Transformer Model for Retrosynthesis

RDChiral: An RDKit Wrapper for Handling Stereochemistry in Retrosynthetic Template Extraction and Application

Enhancing Retrosynthetic Reaction Prediction with Deep Learning Using Multiscale Reaction Classification

Molecular Transformer: A Model for Uncertainty-Calibrated Chemical Reaction Prediction

A probabilistic molecular fingerprint for big data settings

A graph-convolutional neural network model for the prediction of chemical reactivity

Machine Learning in Drug Discovery

Large-scale comparison of machine learning methods for drug target prediction on ChEMBL

Unsupervised Feature Learning via Non-parametric Instance Discrimination

Machine Learning in Computer-Aided Synthesis Planning.

Computer-Assisted Retrosynthesis Based on Molecular Similarity

Decoupled Weight Decay Regularization

Planning chemical syntheses with deep neural networks and symbolic AI

Attention is All you Need

Retrosynthetic Reaction Prediction Using Neural Sequence-to-Sequence Models

Neural-Symbolic Machine Learning for Retrosynthesis and Reaction Prediction.

Diversity in recommender systems - A survey

Neural Message Passing for Quantum Chemistry

Linking the Neural Machine Translation and the Prediction of Organic Chemistry Reactions

What's What: The (Nearly) Definitive Guide to Reaction Role Assignment

Neural Networks for the Prediction of Organic Chemistry Reactions

Layer Normalization

Dense Associative Memory for Pattern Recognition

Computer-Assisted Synthetic Planning: The End of the Beginning.

Challenges and opportunities in computer-aided molecular design

Neural Turing Machines

Sequence to Sequence Learning with Neural Networks

Extraction of chemical structures and reactions from the literature

Extended-Connectivity Fingerprints

Dimensionality Reduction by Learning an Invariant Mapping

A guide to drug discovery: The role of the medicinal chemist in drug discovery — then and now

Storage Capacity of Kernel Associative Memories

Capacities of multiconnected memory models

SMILES, a chemical language and information system. 1. Introduction to methodology and encoding rules

Storage capacity of generalized networks.

Computer-aided molecular design.

Multiconnected neural network models

Number of stable points for spin-glasses and neural networks of higher orders.

Retro: Learning Retrosynthetic Planning with Neural Guided A Search