arXiv subjects

Zixiang Lu

Publications and source records attributed to Zixiang Lu.

6 recordsLinked to original sources

arXiv2026

Selecting optimal unrestricted Hartree-Fock trial wavefunctions for phaseless auxiliary-field quantum Monte Carlo: Accuracy and limitations in modeling three iron-sulfur clusters

Phaseless auxiliary-field quantum Monte Carlo (ph-AFQMC) has emerged as a promising electronic structure method for correlated electronic systems. However, the quality of its predictions depends critically on the choice of trial wavefunction, and it is not obvious how to make an optimal choice especially for strongly correlated states of large systems. Mean-field wavefunctions are compelling trial wavefunction candidates as they map directly to chemical concepts and can be obtained with $O(N^4)$ cost. Yet in the strongly correlated regime one faces a symmetry dilemma and the existence of multiple nearly-degenerate solutions. In this work we investigate active space models of [2Fe-2S]$^{2+}$, mixed-valent [4Fe-4S]$^{2+}$, and [4Fe-4S]$^{4+}$ and explore the sensitivity of ph-AFQMC to the choice of unrestricted Hartree-Fock trial wavefunction. We find that chemical properties and physical symmetries, rather than the variational energy, ought to guide the choice of mean-field trial for ph-AFQMC (or reference state for coupled cluster models), and show that surprisingly accurate ground-state energies for these systems can be obtained. However, in all cases we find a rapidly vanishing overlap between the stochastic wavefunction and the UHF trial, indicating that the trials are suboptimal importance functions. By analogy to a similar situation in the stretched helium dimer cation, we show how this sampling bias pushes ph-AFQMC towards artificially negative energies, which evidently can be compensated for by the phaseless bias in certain cases.

physics.chem-ph↗

arXiv2025

Rethinking LLM Parametric Knowledge as Post-retrieval Confidence for Dynamic Retrieval and Reranking

Large Language Models (LLMs) often generate inaccurate responses (hallucinations) when faced with questions beyond their knowledge scope. Retrieval-Augmented Generation (RAG) addresses this by leveraging external knowledge, but a critical challenge remains: determining whether retrieved contexts effectively enhance the model`s ability to answer specific queries. This challenge underscores the importance of knowledge boundary awareness, which current methods-relying on discrete labels or limited signals-fail to address adequately, as they overlook the rich information in LLMs` continuous internal hidden states. To tackle this, we propose a novel post-retrieval knowledge filtering approach. First, we construct a confidence detection model based on LLMs` internal hidden states to quantify how retrieved contexts enhance the model`s confidence. Using this model, we build a preference dataset (NQ_Rerank) to fine-tune a reranker, enabling it to prioritize contexts preferred by the downstream LLM during reranking. Additionally, we introduce Confidence-Based Dynamic Retrieval (CBDR), which adaptively triggers retrieval based on the LLM`s initial confidence in the original question, reducing knowledge conflicts and improving efficiency. Experimental results demonstrate significant improvements in accuracy for context screening and end-to-end RAG performance, along with a notable reduction in retrieval costs while maintaining competitive accuracy.

cs.IR↗

arXiv2025

MolReasoner: Toward Effective and Interpretable Reasoning for Molecular LLMs

Large Language Models (LLMs) have shown impressive performance across various domains, but their ability to perform molecular reasoning remains underexplored. Existing methods mostly rely on general-purpose prompting, which lacks domain-specific molecular semantics, or fine-tuning, which faces challenges in interpretability and reasoning depth, often leading to structural and textual hallucinations. To address these issues, we introduce MolReasoner, a two-stage framework that transitions LLMs from memorization to high-fidelity chemical reasoning. In the Mol-SFT stage, knowledge-enhanced Chain-of-Thought (CoT) data provides a strong foundation, while the Mol-RL stage refines reasoning using a novel, task-adaptive reward system to mitigate hallucinations. Extensive evaluations demonstrate that MolReasoner significantly outperforms a wide range of strong baselines in both molecule generation and captioning tasks. Further analyses highlight the framework's synergistic design and its ability to produce more interpretable outputs. Our work presents a principled and effective new approach for advancing high-fidelity molecular reasoning.

cs.LG↗

arXiv2025

Implementing advanced trial wave functions in fermion quantum Monte Carlo via stochastic sampling

We introduce an efficient approach to implement correlated many-body trial wave functions in auxiliary-field quantum Monte Carlo (AFQMC). To control the sign/phase problem in AFQMC, a constraint is derived from an exact gauge condition but is typically imposed approximately through a trial wave function or trial density matrix, whose quality can affect the accuracy of the method. Furthermore, the trial wave function can also affect the efficiency through importance sampling. The most natural form of the trial wave function has been single Slater determinants or their linear combinations. More sophisticated forms, for example, with the inclusion of a Jastrow factor or other explicit correlations, have been challenging to use and their implementation is often assumed to require a quantum computer. In this work, we demonstrate that a large class of correlated wave functions, written in the general form of multi-dimensional integrals over hidden or auxiliary variables times Slater determinants, can be implemented as trial wave function by coupling the random walkers to a generalized Metropolis sampling. We discuss the fidelity of AFQMC with stochastically sampled trial wave functions, which are relevant to both quantum and classical algorithms. We illustrate the method and show that an efficient implementation can be achieved which preserves the low-polynomial computational scaling of AFQMC. We test our method in molecules under bond stretching and in transition metal diatomics. Significant improvements are seen in both accuracy and efficiency over typical trial wave functions, and the method yields total ground-state energies systematically within chemical accuracy. The method can be useful for incorporating other advanced wave functions, for example, neural quantum state wave functions optimized from machine learning techniques, or for other forms of fermion quantum Monte Carlo.

cond-mat.str-el↗

arXiv2024

Symmetry enforced solution of the many-body Schr\"odinger equation with deep neural network

The integration of deep neural networks with the Variational Monte Carlo (VMC) method has marked a significant advancement in solving the Schr\"odinger equation. In this work, we enforce spin symmetry in the neural network-based VMC calculation with modified optimization target. Our method is designed to solve for the ground state and multiple excited states with target spin symmetry at a low computational cost. It predicts accurate energies while maintaining the correct symmetry in strongly correlated systems, even in cases where different spin states are nearly degenerate. Our approach also excels at spin-gap calculations, including the singlet-triplet gap in biradical systems, which is of high interest in photochemistry. Overall, this work establishes a robust framework for efficiently calculating various quantum states with specific spin symmetry in correlated systems, paving the way for novel discoveries in quantum science.

physics.chem-ph↗

arXiv2023

Penalty and auxiliary wave function methods for electronic Excitation in neural network variational Monte Carlo

This study explores the application of neural network variational Monte Carlo (NN-VMC) for the computation of low-lying excited states in molecular systems. Our focus lies on the implementation and evaluation of two distinct methodologies, the penalty method and a novel modification of the auxiliary wave function (AW) method, within the framework of the FermiNet-based NN-VMC package. Importantly, this specific application has not been previously reported.Our investigation advocates for the efficacy of the modified AW method, emphasizing its superior robustness when compared to the penalty method. This methodological advancement introduces a valuable tool for the scientific community, offering a distinctive approach to target low-lying excited states. We anticipate that the modified AW method will garner interest within the research community, serving as a complementary and robust alternative to existing techniques. Moreover, this contribution enriches the ongoing development of various neural network ansatz, further expanding the toolkit available for the accurate exploration of excited states in molecular systems.

physics.comp-ph↗