Generalizable, Fast, and Accurate DeepQSPR with fastprop Part 1: Framework and Benchmarks

2024-04-02

arXiv.org

AI Summary

Traditional QSPR methods rely on expert-crafted molecular descriptors, suffering from poor generalizability and limited adaptability. This paper proposes *fastprop*, a lightweight QSPR framework that combines physicochemically inspired handcrafted features with a multilayer perceptron (MLP), deliberately avoiding complex graph neural networks (GNNs) or Transformers for end-to-end representation learning—thereby balancing interpretability, predictive accuracy, and computational efficiency. Through systematic hyperparameter optimization and cross-dataset standardized preprocessing, *fastprop* achieves significantly lower mean absolute error (MAE) and root-mean-square error (RMSE) than leading models—including GCN, MPNN, and D-MPNN—across 23 molecular property prediction benchmarks. It reduces training time by over 90% and accelerates inference by one to two orders of magnitude. The implementation and pretrained models are publicly available.

Technology Category

• Machine Learning: Feature Construction/Reformulation
• Knowledge Representation and Reasoning: Computational Complexity of Reasoning
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Quantitative Structure Property Relationship studies aim to define a mapping between molecular structure and arbitrary quantities of interest. This was historically accomplished via the development of descriptors which requires significant domain expertise and struggles to generalize. Thus the field has morphed into Molecular Property Prediction and been given over to learned representations which are highly generalizable. The paper introduces fastprop, a DeepQSPR framework which uses a cogent set of molecular level descriptors to meet and exceed the performance of learned representations on diverse datasets in dramatically less time. fastprop is freely available on github at github.com/JacksonBurns/fastprop.

Problem

• Quantitative Structure-Property Relationships
• Descriptor Engineering
• Predictive Modeling

Innovation

• fastprop
• learning descriptors
• DeepQSPR framework