Large Language Models Can Help Mitigate Barren Plateaus in Quantum Neural Networks
Abstract
In the era of noisy intermediate-scale quantum (NISQ) computing, Quantum Neural Networks (QNNs) have emerged as a promising approach for various applications, yet their training is often hindered by barren plateaus (BPs), where gradient variance vanishes exponentially as the qubit size increases. Most initialization-based mitigation strategies rely heavily on pre-designed static parameter distributions, thereby lacking adaptability to diverse model sizes or data conditions. To address these limitations, we propose AdaInit, a foundational framework that leverages large language models with the submartingale property to iteratively synthesize initial parameters for QNNs that yield non-negligible gradient variance, thereby mitigating BPs. Unlike conventional one-shot initialization methods, AdaInit adaptively explores the parameter space by incorporating dataset characteristics and gradient feedback, with theoretical guarantees of convergence to finding a set of effective initial parameters for QNNs. We provide rigorous theoretical analyses of the submartingale-based process and empirically validate that AdaInit consistently outperforms existing initialization methods in maintaining higher gradient variance across various QNN scales. We believe this work may initiate a new avenue to mitigate BPs.
Cite
@article{arxiv.2502.13166,
title = {Large Language Models Can Help Mitigate Barren Plateaus in Quantum Neural Networks},
author = {Jun Zhuang and Chaowen Guan},
journal= {arXiv preprint arXiv:2502.13166},
year = {2026}
}
Comments
[ACL'26 Findings] TL;DR: We propose a new LLM-driven submartingale-based framework that adaptively generates effective initial parameters for quantum neural networks to mitigate barren plateaus by leveraging LLMs with the submartingale property