FusionRetro: Molecule Representation Fusion via In-Context Learning for Retrosynthetic Planning

Songtao Liu, Zhengkai Tu, Minkai Xu, Zuobai Zhang, Lu Lin, Rex Ying, Jian Tang, Peilin Zhao, Dinghao Wu
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:22028-22041, 2023.

Abstract

Retrosynthetic planning aims to devise a complete multi-step synthetic route from starting materials to a target molecule. Current strategies use a decoupled approach of single-step retrosynthesis models and search algorithms, taking only the product as the input to predict the reactants for each planning step and ignoring valuable context information along the synthetic route. In this work, we propose a novel framework that utilizes context information for improved retrosynthetic planning. We view synthetic routes as reaction graphs and propose to incorporate context through three principled steps: encode molecules into embeddings, aggregate information over routes, and readout to predict reactants. Our approach is the first attempt to utilize in-context learning for retrosynthesis prediction in retrosynthetic planning. The entire framework can be efficiently optimized in an end-to-end fashion and produce more practical and accurate predictions. Comprehensive experiments demonstrate that by fusing in the context information over routes, our model significantly improves the performance of retrosynthetic planning over baselines that are not context-aware, especially for long synthetic routes. Code is available at https://github.com/SongtaoLiu0823/FusionRetro.

Cite this Paper

BibTeX
@InProceedings{pmlr-v202-liu23ah, title = {{F}usion{R}etro: Molecule Representation Fusion via In-Context Learning for Retrosynthetic Planning}, author = {Liu, Songtao and Tu, Zhengkai and Xu, Minkai and Zhang, Zuobai and Lin, Lu and Ying, Rex and Tang, Jian and Zhao, Peilin and Wu, Dinghao}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {22028--22041}, year = {2023}, editor = {Krause, Andreas and Brunskill, Emma and Cho, Kyunghyun and Engelhardt, Barbara and Sabato, Sivan and Scarlett, Jonathan}, volume = {202}, series = {Proceedings of Machine Learning Research}, month = {23--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v202/liu23ah/liu23ah.pdf}, url = {https://proceedings.mlr.press/v202/liu23ah.html}, abstract = {Retrosynthetic planning aims to devise a complete multi-step synthetic route from starting materials to a target molecule. Current strategies use a decoupled approach of single-step retrosynthesis models and search algorithms, taking only the product as the input to predict the reactants for each planning step and ignoring valuable context information along the synthetic route. In this work, we propose a novel framework that utilizes context information for improved retrosynthetic planning. We view synthetic routes as reaction graphs and propose to incorporate context through three principled steps: encode molecules into embeddings, aggregate information over routes, and readout to predict reactants. Our approach is the first attempt to utilize in-context learning for retrosynthesis prediction in retrosynthetic planning. The entire framework can be efficiently optimized in an end-to-end fashion and produce more practical and accurate predictions. Comprehensive experiments demonstrate that by fusing in the context information over routes, our model significantly improves the performance of retrosynthetic planning over baselines that are not context-aware, especially for long synthetic routes. Code is available at https://github.com/SongtaoLiu0823/FusionRetro.} }
Endnote
%0 Conference Paper %T FusionRetro: Molecule Representation Fusion via In-Context Learning for Retrosynthetic Planning %A Songtao Liu %A Zhengkai Tu %A Minkai Xu %A Zuobai Zhang %A Lu Lin %A Rex Ying %A Jian Tang %A Peilin Zhao %A Dinghao Wu %B Proceedings of the 40th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2023 %E Andreas Krause %E Emma Brunskill %E Kyunghyun Cho %E Barbara Engelhardt %E Sivan Sabato %E Jonathan Scarlett %F pmlr-v202-liu23ah %I PMLR %P 22028--22041 %U https://proceedings.mlr.press/v202/liu23ah.html %V 202 %X Retrosynthetic planning aims to devise a complete multi-step synthetic route from starting materials to a target molecule. Current strategies use a decoupled approach of single-step retrosynthesis models and search algorithms, taking only the product as the input to predict the reactants for each planning step and ignoring valuable context information along the synthetic route. In this work, we propose a novel framework that utilizes context information for improved retrosynthetic planning. We view synthetic routes as reaction graphs and propose to incorporate context through three principled steps: encode molecules into embeddings, aggregate information over routes, and readout to predict reactants. Our approach is the first attempt to utilize in-context learning for retrosynthesis prediction in retrosynthetic planning. The entire framework can be efficiently optimized in an end-to-end fashion and produce more practical and accurate predictions. Comprehensive experiments demonstrate that by fusing in the context information over routes, our model significantly improves the performance of retrosynthetic planning over baselines that are not context-aware, especially for long synthetic routes. Code is available at https://github.com/SongtaoLiu0823/FusionRetro.
APA
Liu, S., Tu, Z., Xu, M., Zhang, Z., Lin, L., Ying, R., Tang, J., Zhao, P. & Wu, D.. (2023). FusionRetro: Molecule Representation Fusion via In-Context Learning for Retrosynthetic Planning. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:22028-22041 Available from https://proceedings.mlr.press/v202/liu23ah.html.

Related Material