Improving Few- and Zero-Shot Reaction Template Prediction Using Modern Hopfield Networks

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.

[1]  Christopher D. Manning,et al.  Contrastive Learning of Medical Visual Representations from Paired Images and Text , 2020, MLHC.

[2]  Jinwoo Shin,et al.  RetCL: A Selection-based Approach for Retrosynthesis via Contrastive Learning , 2021, IJCAI.

[3]  Huanxiang Liu,et al.  RetroPrime: A Diverse, Plausible and Transformer-based Method for Single-Step Retrosynthesis Predictions , 2021 .

[4]  Venkat Venkatasubramanian,et al.  Retrosynthesis prediction using grammar-based neural machine translation: An information-theoretic approach , 2021, Comput. Chem. Eng..

[5]  Alan Hanjalic,et al.  Accuracy-diversity trade-off in recommender systems via graph convolutions , 2021, Inf. Process. Manag..

[6]  Ilya Sutskever,et al.  Learning Transferable Visual Models From Natural Language Supervision , 2021, ICML.

[7]  Takashi Ishida,et al.  Single-Step Retrosynthesis Prediction Based on the Identification of Potential Disconnection Sites Using Molecular Substructure Fingerprints , 2021, J. Chem. Inf. Model..

[8]  Juyong Lee,et al.  Substructure-based neural machine translation for retrosynthetic prediction , 2020, Journal of Cheminformatics.

[9]  Stanislaw Jastrzebski,et al.  Molecule Edit Graph Attention Network: Modeling Chemical Reactions as Sequences of Graph Edits , 2020, J. Chem. Inf. Model..

[10]  O. Engkvist,et al.  Artificial applicability labels for improving policies in retrosynthesis prediction , 2020, Mach. Learn. Sci. Technol..

[11]  Matt J. Kusner,et al.  Barking up the right tree: an approach to search over molecule synthesis DAGs , 2020, NeurIPS.

[12]  Xianzhen Wang,et al.  Interpretable machine-learning strategy for soft-magnetic property and thermal stability in Fe-based metallic glasses , 2020, npj Computational Materials.

[13]  Katsuhiko Ishiguro,et al.  Data Transfer Approaches to Improve Seq-to-Seq Retrosynthesis , 2020, ArXiv.

[14]  Marwin H. S. Segler,et al.  Machine learning the ropes: principles, applications and directions in synthetic chemistry. , 2020, Chemical Society reviews.

[15]  Le Song,et al.  Retro*: Learning Retrosynthetic Planning with Neural Guided A* Search , 2020, ICML.

[16]  Yang Yu,et al.  RetroXpert: Decompose Retrosynthesis Prediction like a Chemist , 2020, NeurIPS.

[17]  Regina Barzilay,et al.  Learning Graph Models for Template-Free Retrosynthesis , 2020, ArXiv.

[18]  Emmanuel Noutahi,et al.  Molecular Design in Synthetically Accessible Chemical Space via Deep Reinforcement Learning , 2020, ACS omega.

[19]  Yoshua Bengio,et al.  Learning To Navigate The Synthetically Accessible Chemical Space Using Reinforcement Learning , 2020, ICML.

[20]  Geir Kjetil Sandve,et al.  Modern Hopfield Networks and Attention for Immune Repertoire Classification , 2020, bioRxiv.

[21]  Regina Barzilay,et al.  Current and Future Roles of Artificial Intelligence in Medicinal Chemistry Synthesis , 2020, Journal of medicinal chemistry.

[22]  Jian Tang,et al.  A Graph to Graphs Framework for Retrosynthesis Prediction , 2020, ICML.

[23]  Stephen Wu,et al.  A Bayesian algorithm for retrosynthesis , 2020, Journal of chemical information and modeling.

[24]  I. Tetko,et al.  State-of-the-art augmented NLP transformer models for direct and single-step retrosynthesis , 2020, Nature Communications.

[25]  Geoffrey E. Hinton,et al.  A Simple Framework for Contrastive Learning of Visual Representations , 2020, ICML.

[26]  Brian C. Barnes,et al.  Data Augmentation and Pretraining for Template-Based Retrosynthetic Prediction in Computer-Aided Synthesis Planning , 2020, J. Chem. Inf. Model..

[27]  Laurens van der Maaten,et al.  Self-Supervised Learning of Pretext-Invariant Representations , 2019, 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[28]  Ross B. Girshick,et al.  Momentum Contrast for Unsupervised Visual Representation Learning , 2019, 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[29]  Junzhou Huang,et al.  Molecular Graph Enhanced Transformer for Retrosynthesis Prediction , 2019, bioRxiv.

[30]  Yuedong Yang,et al.  Predicting Retrosynthetic Reaction using Self-Corrected Transformer Neural Networks , 2019, J. Chem. Inf. Model..

[31]  Phillip Isola,et al.  Contrastive Multiview Coding , 2019, ECCV.

[32]  Le Song,et al.  Retrosynthesis Prediction with Conditional Graph Logic Network , 2020, NeurIPS.

[33]  Marwin H. S. Segler,et al.  World Programs for Model-Based Learning and Planning in Compositional State and Action Spaces , 2019, ArXiv.

[34]  Natalia Gimelshein,et al.  PyTorch: An Imperative Style, High-Performance Deep Learning Library , 2019, NeurIPS.

[35]  Sepp Hochreiter,et al.  On failure modes in molecule generation and optimization. , 2019, Drug discovery today. Technologies.

[36]  Regina Barzilay,et al.  Learning to Make Generalizable and Diverse Predictions for Retrosynthesis , 2019, ArXiv.

[37]  R. Kojima,et al.  Prediction and Interpretable Visualization of Retrosynthetic Reactions Using Graph Convolutional Networks , 2019, J. Chem. Inf. Model..

[38]  Matt J. Kusner,et al.  A Model to Search for Synthesizable Molecules , 2019, NeurIPS.

[39]  Igor V. Tetko,et al.  A Transformer Model for Retrosynthesis , 2019, ICANN.

[40]  Connor W. Coley,et al.  RDChiral: An RDKit Wrapper for Handling Stereochemistry in Retrosynthetic Template Extraction and Application , 2019, J. Chem. Inf. Model..

[41]  Nicholas A Cilfone,et al.  Enhancing Retrosynthetic Reaction Prediction with Deep Learning Using Multiscale Reaction Classification , 2019, J. Chem. Inf. Model..

[42]  Connor W. Coley,et al.  A graph-convolutional neural network model for the prediction of chemical reactivity† †Electronic supplementary information (ESI) available: Additional model and dataset details, results, discussion, and ref. 38 and 39. See DOI: 10.1039/c8sc04228d , 2018, Chemical science.

[43]  Matthias Rarey,et al.  Machine Learning in Drug Discovery. , 2018, Journal of chemical information and modeling.

[44]  Frank Hutter,et al.  Decoupled Weight Decay Regularization , 2017, ICLR.

[45]  Constantine Bekas,et al.  Molecular Transformer for Chemical Reaction Prediction and Uncertainty Estimation , 2018, ArXiv.

[46]  Jean-Louis Reymond,et al.  A probabilistic molecular fingerprint for big data settings , 2018, Journal of Cheminformatics.

[47]  Hugo Ceulemans,et al.  Large-scale comparison of machine learning methods for drug target prediction on ChEMBL† †Electronic supplementary information (ESI) available: Overview, Data Collection and Clustering, Methods, Results, Appendix. See DOI: 10.1039/c8sc00148k , 2018, Chemical science.

[48]  Stella X. Yu,et al.  Unsupervised Feature Learning via Non-parametric Instance Discrimination , 2018, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[49]  Connor W. Coley,et al.  Machine Learning in Computer-Aided Synthesis Planning. , 2018, Accounts of chemical research.

[50]  Constantine Bekas,et al.  “Found in Translation”: predicting outcomes of complex organic chemistry reactions using neural sequence-to-sequence models† †Electronic supplementary information (ESI) available: Time-split test set and example predictions, together with attention weights, confidence and token probabilities. See DO , 2017, Chemical science.

[51]  William H. Green,et al.  Computer-Assisted Retrosynthesis Based on Molecular Similarity , 2017, ACS central science.

[52]  Mike Preuss,et al.  Learning to Plan Chemical Syntheses , 2017, ArXiv.

[53]  Lukasz Kaiser,et al.  Attention is All you Need , 2017, NIPS.

[54]  Bowen Liu,et al.  Retrosynthetic Reaction Prediction Using Neural Sequence-to-Sequence Models , 2017, ACS central science.

[55]  Marwin H. S. Segler,et al.  Neural-Symbolic Machine Learning for Retrosynthesis and Reaction Prediction. , 2017, Chemistry.

[56]  Matevz Kunaver,et al.  Diversity in recommender systems - A survey , 2017, Knowl. Based Syst..

[57]  Samuel S. Schoenholz,et al.  Neural Message Passing for Quantum Chemistry , 2017, ICML.

[58]  Juno Nam,et al.  Linking the Neural Machine Translation and the Prediction of Organic Chemistry Reactions , 2016, ArXiv.

[59]  Gregory A Landrum,et al.  What's What: The (Nearly) Definitive Guide to Reaction Role Assignment , 2016, J. Chem. Inf. Model..

[60]  Alán Aspuru-Guzik,et al.  Neural Networks for the Prediction of Organic Chemistry Reactions , 2016, ACS central science.

[61]  John J. Hopfield,et al.  Dense Associative Memory for Pattern Recognition , 2016, NIPS.

[62]  Piotr Dittwald,et al.  Computer-Assisted Synthetic Planning: The End of the Beginning. , 2016, Angewandte Chemie.

[63]  Nishanth G. Chemmangattuvalappil,et al.  Challenges and opportunities in computer-aided molecular design , 2015, Comput. Chem. Eng..

[64]  Quoc V. Le,et al.  Sequence to Sequence Learning with Neural Networks , 2014, NIPS.

[65]  Daniel M. Lowe.  Extraction of chemical structures and reactions from the literature , 2012 .

[66]  David Rogers,et al.  Extended-Connectivity Fingerprints , 2010, J. Chem. Inf. Model..

[67]  Yann LeCun,et al.  Dimensionality Reduction by Learning an Invariant Mapping , 2006, 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06).

[68]  John A. Lowe,et al.  A guide to drug discovery: The role of the medicinal chemist in drug discovery — then and now , 2004, Nature Reviews Drug Discovery.

[69]  Heinrich Niemann,et al.  Storage Capacity of Kernel Associative Memories , 2002, ICANN.

[70]  M. Usher,et al.  Capacities of multiconnected memory models , 1988 .

[71]  David Weininger,et al.  SMILES, a chemical language and information system. 1. Introduction to methodology and encoding rules , 1988, J. Chem. Inf. Comput. Sci..

[72]  Abbott,et al.  Storage capacity of generalized networks. , 1987, Physical review. A, General physics.

[73]  J A McCammon,et al.  Computer-aided molecular design. , 1987, Science.

[74]  E. Gardner.  Multiconnected neural network models , 1987 .

[75]  Baldi,et al.  Number of stable points for spin-glasses and neural networks of higher orders. , 1987, Physical review letters.

[76]  C. L. Giles,et al.  High order correlation model for associative memory , 1987 .

[77]  Demetri Psaltis,et al.  Nonlinear discriminant functions and associative memories , 1987 .

[78]  J J Hopfield,et al.  Neural networks and physical systems with emergent collective computational abilities. , 1982, Proceedings of the National Academy of Sciences of the United States of America.

[79]  E J Corey,et al.  Computer-assisted design of complex organic syntheses. , 1969, Science.