DeepAI AI Chat
Log In Sign Up

Retrosynthetic Planning with Dual Value Networks

by   Guoqing Liu, et al.

Retrosynthesis, which aims to find a route to synthesize a target molecule from commercially available starting materials, is a critical task in drug discovery and materials design. Recently, the combination of ML-based single-step reaction predictors with multi-step planners has led to promising results. However, the single-step predictors are mostly trained offline to optimize the single-step accuracy, without considering complete routes. Here, we leverage reinforcement learning (RL) to improve the single-step predictor, by using a tree-shaped MDP to optimize complete routes while retaining single-step accuracy. Desirable routes should be both synthesizable and of low cost. We propose an online training algorithm, called Planning with Dual Value Networks (PDVN), in which two value networks predict the synthesizability and cost of molecules, respectively. To maintain the single-step accuracy, we design a two-branch network structure for the single-step predictor. On the widely-used USPTO dataset, our PDVN algorithm improves the search success rate of existing multi-step planners (e.g., increasing the success rate from 85.79 to 98.95 solving 99.47 synthesis routes (e.g., reducing the average route length from 5.76 to 4.83 for Retro*, and from 5.63 to 4.78 for RetroGraph).


page 1

page 2

page 3

page 4


Metro: Memory-Enhanced Transformer for Retrosynthetic Planning via Reaction Tree

Retrosynthetic planning plays a critical role in drug discovery and orga...

Learning retrosynthetic planning through self-play

The problem of retrosynthetic planning can be framed as one player game,...

Modern Hopfield Networks for Few- and Zero-Shot Reaction Prediction

An essential step in the discovery of new drugs and materials is the syn...

Branch Prediction as a Reinforcement Learning Problem: Why, How and Case Studies

Recent years have seen stagnating improvements to branch predictor (BP) ...

Self-Improved Retrosynthetic Planning

Retrosynthetic planning is a fundamental problem in chemistry for findin...

Learning to Plan Chemical Syntheses

From medicines to materials, small organic molecules are indispensable f...

Improving Branch Prediction By Modeling Global History with Convolutional Neural Networks

CPU branch prediction has hit a wall--existing techniques achieve near-p...