An ab initio foundation model of wavefunctions that accurately describes chemical bond breaking

2025-06-24

AI Summary

Chemical bond dissociation often exhibits strong multireference character, rendering conventional multireference methods computationally expensive and lacking transferability across molecular systems. To address this, we propose Orbformer—the first deep neural network–based wavefunction pretraining framework that integrates quantum Monte Carlo with the large-scale pretraining–fine-tuning paradigm, trained on 22,000 molecular structures spanning equilibrium and dissociation geometries. Orbformer learns transferable, general-purpose electronic structure representations applicable across diverse molecules, drastically reducing computational cost. On multiple bond dissociation pathways and Diels–Alder reactions, it achieves chemical accuracy (≤1 kcal/mol), matching the accuracy of state-of-the-art multireference methods while offering orders-of-magnitude speedup. This work establishes a scalable, efficient machine learning alternative for multireference problems in quantum chemistry.

Technology Category

• Machine Learning: Quantum Machine Learning
• Search and Optimization: Learning to Search
• Multiagent Systems: Multiagent Learning

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Economics, Online Markets and Human Computation: Architectures and workflows that use LLMs for crowd work
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Reliable description of bond breaking remains a major challenge for quantum chemistry due to the multireferential character of the electronic structure in dissociating species. Multireferential methods in particular suffer from large computational cost, which under the normal paradigm has to be paid anew for each system at a full price, ignoring commonalities in electronic structure across molecules. Quantum Monte Carlo with deep neural networks (deep QMC) uniquely offers to exploit such commonalities by pretraining transferable wavefunction models, but all such attempts were so far limited in scope. Here, we bring this new paradigm to fruition with Orbformer, a novel transferable wavefunction model pretrained on 22,000 equilibrium and dissociating structures that can be fine-tuned on unseen molecules reaching an accuracy-cost ratio rivalling classical multireferential methods. On established benchmarks as well as more challenging bond dissociations and Diels-Alder reactions, Orbformer is the only method that consistently converges to chemical accuracy (1 kcal/mol). This work turns the idea of amortizing the cost of solving the Schrödinger equation over many molecules into a practical approach in quantum chemistry.

Problem

• Accurately describing chemical bond breaking in quantum chemistry
• Reducing computational cost of multireferential methods for dissociating species
• Developing transferable wavefunction models for diverse molecular structures

Innovation

• Pretrained transferable wavefunction model Orbformer
• Quantum Monte Carlo with deep neural networks
• Amortizes cost over 22,000 molecular structures