Skip to main content
QUICK REVIEW

[论文解读] Data-efficient surrogate modeling of spectral functions using Gaussian processes: An application to the $t$-$t'$-$t''$-$J$ model

Sanket Jantre, Nathan M. Urban|arXiv (Cornell University)|Mar 13, 2026
Machine Learning in Materials Science被引用 0
一句话总结

一个数据高效的前向代理,用深核高斯过程和稀疏变分推断,在 t–t′–t″–J 模型的光谱函数上表现优于数据削减的 FFNN 基线,并接近全数据表现。

ABSTRACT

Spectral functions encode key many-body information but are costly to compute with high fidelity. Machine-learning surrogates have emerged as a powerful alternative, yet many approaches require large training datasets. We develop a data-efficient surrogate for spectral functions using the $t$-$t'$-$t''$-$J$ model, which describes the motion of a hole in a quantum antiferromagnet. Using $\sim$ 10$^5$ self-consistent Born approximation-based spectra from Lee, Carbone and Yin (Phys. Rev. B 107, 205132 (2023)), we train a deep-kernel Gaussian process surrogate model with sparse variational inference (DKL-SVGP) using only 10% of the available training spectra. We benchmark against feed-forward neural networks (FFNN) trained on the same reduced subset and on the full dataset. The proposed DKL-SVGP model consistently outperforms the reduced-data FFNN and, despite using only 10% of the training spectra, achieves spectrum-wise errors within the same order-of-magnitude as the full-data FFNN baseline. Worst-tail diagnostics show improved fidelity on difficult spectra, while peak-level analysis indicates that DKL-SVGP recovers dominant peak heights with comparable accuracy and improves peak-location agreement under a matched-peak evaluation that mitigates rare peak-swapping cases. Overall, these results highlight GP-based surrogates as a competitive and data-efficient approach for spectral-function prediction in scarce-data regimes.

研究动机与目标

  • 在有限训练数据下使光谱函数代理工作良好。
  • 开发将哈密顿量参数与态密度 A(ω) 联系起来的高效前向模型。
  • 证明深核 SVGP 代理在数据削减的基线之上表现更好。
  • 评估光谱层级和峰值层级的保真度,包括最坏尾部和匹配峰的诊断。
  • 强调在光谱问题中的主动学习和反演建模的意义。

提出的方法

  • 将 A(ω) 视为参数 x=(t', t'', J) 与能量 ω 的函数;在固定 ω 网格上对光谱进行训练。
  • 使用带稀疏变分高斯过程的深核学习(DKL)代理(DKL-SVGP),诱导点 M=1536。
  • 使用轻量特征网络 φθ,结合参数网络 hx 与傅里叶 ω-主干,输入到具有灵活核的 GP。
  • 对联合输入 zjoint=[φθ(s), ω̃] 建模混合核 k(s,z'),通过在学习空间中的 Matérn-3/2 核耦合学习到的特征和 ω,以及针对 ω 的谱混合核。
  • 预处理:对 x 进行标准化,对 ω 进行重新缩放,对 A(ω) 进行对数标准化,并在 LCY 数据集的 10% 子集上训练;将预测均值作为代理输出。
  • 与在相同 10% 子集和全数据上训练的 FFNN 基线进行比较;附录中包含 KRR 和 kNN 基线。
Figure 1: Illustration of our deep kernel Gaussian process surrogate model trained with stochastic variational inference and applied to the forward problem of predicting a DOS given Hamiltonian parameters $x=(t^{\prime},t^{\prime\prime},J)$ and energy-grid points $\omega$ . We first standardize $x\t
Figure 1: Illustration of our deep kernel Gaussian process surrogate model trained with stochastic variational inference and applied to the forward problem of predicting a DOS given Hamiltonian parameters $x=(t^{\prime},t^{\prime\prime},J)$ and energy-grid points $\omega$ . We first standardize $x\t

实验结果

研究问题

  • RQ1GP 基代理在深核学习下,是否能够用仅 10% 的训练数据实现对 A(ω) 的具有竞争力的精度?
  • RQ2DKL-SVGP 相对于削减数据的 FFNN 和全数据 FFNN,在光谱层级和峰值层级的保真度如何?
  • RQ3该代理在困难光谱(最坏尾部)上是否保持保真,并在匹配峰评估下更好地恢复峰高和峰位?

主要发现

  • 在 10% 数据下训练的 DKL-SVGP 相比同样子集训练的 FFNN,测试行均方误差(mean row-RMSE)降低约 42%。
  • FFNN(全数据)与 DKL-SVGP 的测试行均方误差分别为 1.93e-3 和 4.54e-3(DKL-SVGP 相当于全数据 FFNN 的约 2.4 倍误差)。
  • 测试集归一化 RMSE(NRMSE):FFNN(全数据) 1.365e-2;DKL-SVGP 3.369e-2(子集),显示显著的数据高效收益。
  • 最坏尾部诊断显示 DKL-SVGP 在困难光谱上更好地保持主导光谱结构,相较于削减数据的 FFNN。
  • 在匹配峰评估下的峰值层次保真:峰高相当,峰位方面 DKL-SVGP 略更准确(RMSE 0.021 与 FFNN 子集 0.028)。
  • 总体而言,基于 GP 的代理在数据稀缺情境下为光谱建模提供了强大的数据效率与准确性。
Figure 2: Worst-tail diagnostic on the held-out test dataset. Test spectra are ranked from worst to best by a reference error score (here, the FFNN row-RMSE over test dataset), and representative spectra at percentiles $\{0,2,4,6,8,10\}\%$ within this worst tail are shown. Each panel overlays the gr
Figure 2: Worst-tail diagnostic on the held-out test dataset. Test spectra are ranked from worst to best by a reference error score (here, the FFNN row-RMSE over test dataset), and representative spectra at percentiles $\{0,2,4,6,8,10\}\%$ within this worst tail are shown. Each panel overlays the gr

更好的研究,从现在开始

从阅读论文到最终审阅,大幅缩短您的研究时间。

无需绑定信用卡

本解读由 AI 生成,并经人工编辑审核。