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

Abstract

Organic synthesis is one of the key stumbling blocks in medicinal chemistry. A necessary yet unsolved step in planning synthesis is solving the forward problem: given reactants and reagents, predict the products. Similar to other work, we treat reaction prediction as a machine translation problem between SMILES strings of reactants-reagents and the products. We show that a multi-head attention Molecular Transformer model outperforms all algorithms in the literature, achieving a top-1 accuracy above 90% on a common benchmark dataset. Our algorithm requires no handcrafted rules, and accurately predicts subtle chemical transformations. Crucially, our model can accurately estimate its own uncertainty, with an uncertainty score that is 89% accurate in terms of classifying whether a prediction is correct. Furthermore, we show that the model is able to handle inputs without reactant-reagent split and including stereochemistry, which makes our method universally applicable.

Machine Learning for Molecules and Materials workshop, NeurIPS 2018 / Platform: https://rxn.res.ibm.com

References in corpus (5)

• Automatic chemical design using a data-driven continuous representation of molecules
• Learning to Plan Chemical Syntheses
• Deep Reinforcement Learning for De-Novo Drug Design
• Neural networks for the prediction organic chemistry reactions
• Conditional molecular design with deep generative models

Cited by in corpus (42)

• Molecular Contrastive Learning of Representations via Graph Neural Networks
• Combining Machine Learning and Computational Chemistry for Predictive Insights Into Chemical Systems
• State-of-the-Art Augmented NLP Transformer models for direct and single-step retrosynthesis
• SELFIES and the future of molecular string representations
• Transformer-CNN: Fast and Reliable tool for QSAR
• polyBERT: A chemical language model to enable fully machine-driven ultrafast polymer informatics
• The Exploration of Chemical Reaction Networks
• TransPolymer: a Transformer-based language model for polymer property predictions
• Transformers in Healthcare: A Survey
• Exploring Chemical Space using Natural Language Processing Methodologies for Drug Discovery
• Regression Transformer: Concurrent sequence regression and generation for molecular language modeling
• Quantum chemistry-augmented neural networks for reactivity prediction: Performance, generalizability and interpretability
• Root-aligned SMILES: A Tight Representation for Chemical Reaction Prediction
• Graph Neural Networks for Temperature-Dependent Activity Coefficient Prediction of Solutes in Ionic Liquids
• SPT-NRTL: A physics-guided machine learning model to predict thermodynamically consistent activity coefficients
• Molecular Design in Synthetically Accessible Chemical Space via Deep Reinforcement Learning
• Learning from flowsheets: A generative transformer model for autocompletion of flowsheets
• Bidirectional Generation of Structure and Properties Through a Single Molecular Foundation Model
• A Bayesian algorithm for retrosynthesis
• Toward automatic generation of control structures for process flow diagrams with large language models
• Developing ChemDFM as a large language foundation model for chemistry
• Scientific intuition inspired by machine learning generated hypotheses
• Navigating chemical reaction space with a steering wheel
• Bridging the Gap between Chemical Reaction Pretraining and Conditional Molecule Generation with a Unified Model
• 3DReact: Geometric deep learning for chemical reactions
• UAlign: Pushing the Limit of Template-free Retrosynthesis Prediction with Unsupervised SMILES Alignment
• Spiers Memorial Lecture: How to do impactful research in artificial intelligence for chemistry and materials science
• A generative model for molecule generation based on chemical reaction trees
• A Large Encoder-Decoder Family of Foundation Models For Chemical Language
• Language models in molecular discovery
• The Rise of Generative AI for Metal-Organic Framework Design and Synthesis
• Graph-to-SFILES: Control structure prediction from process topologies using generative artificial intelligence
• Retro-BLEU: Quantifying Chemical Plausibility of Retrosynthesis Routes through Reaction Template Sequence Analysis
• From predicting to learning dissipation from pair correlations of active liquids
• HiPerRAG: High-Performance Retrieval Augmented Generation for Scientific Insights
• Tokenization for Molecular Foundation Models
• SynRXN: An Open Benchmark and Curated Dataset for Computational Reaction Modeling
• A General-Purpose Framework for Chemical Reaction Representation with Atomic Correspondence and Flexible Condition Adaptation
• Clustering in pure-attention hardmax transformers and its role in sentiment analysis
• Efficient interpolation of molecular properties across chemical compound space with low-dimensional descriptors
• SMILESGNN: Interpretable Clinical Toxicity Prediction via SMILES-Graph Cross-Attention Fusion
• Developing an AI Course for Synthetic Chemistry Students