Machine-learned, finite temperature Fermi-operator expansions suitable for GPUs and AI-hardware

Abstract

We present several finite-temperature recursive Fermi-operator expansion schemes based on the second-order spectral projection (SP2) method. Our approach builds on a previous observation that the electronic structure problem, as formulated through a recursive SP2 expansion, can be mapped onto the architecture of a deep neural network. Using this perspective, we generalize SP2 to finite electronic temperatures by constructing machine learning models that determine optimized recursive expansion coefficients. The same approach is also applied to the prediction of the electronic entropy for fractional occupation numbers. The coefficients are trained for a specified chemical potential and electronic temperature and are not available in closed analytical form. However, by employing an appropriate affine rescaling strategy to the Hamiltonian matrix, we eliminate the need to retrain the model for different temperatures and chemical potentials. Our approach avoids explicit diagonalization and relies solely on highly optimized matrix-matrix multiplication kernels. Compared to state-of-the-art diagonalization, we achieve an order-of-magnitude speedup in the single-particle finite-temperature density matrix calculation for small and moderately sized matrices on modern GPUs and dense matrix multiply units.

Transient Heat Conduction on Quantum Hardware: Replicated Temperature-Field Reconstruction and Shot-Count Sensitivity with a Variational Quantum Linear Solver
Yijin Mao, Yuwen Zhang
arXiv15 Oct 2026PreprintarXiv licenceFull textComputational physics Quantum physics & quantum informationarXiv
Quantum Machine Learning for Cybersecurity Applications: Simulation and Hardware Validation
Zirui Zhu, Zisheng Chen, Xiangyang Li
arXiv26 Sep 2026PreprintCC BY 4.0Full textQuantum physics & quantum informationarXiv
χ‑Operator Geometry of Extremal Black Holes: Spectral Degeneracy, Zero Temperature, and Non‑Zero Entropy
Anatolii Mukha
14 Aug 2026PreprintCC BYFull text12 citedHigh-energy & particle physicsOpenAlex
Testing hardware quantum machine learning claims with a null circuit: fill prediction on Binance futures with 129 qubits of an IBM Heron processor
Muhammad Faryad
arXiv30 Oct 2026PreprintCC BY-NC-SA 4.0Full textQuantum physics & quantum informationarXiv
Finite-temperature operator basis on ℝ³ × S¹ for SMEFT
Joydeep Chakrabortty, Bruno Siqueira Eduardo, Siddhartha Karmakar +1
arXiv28 Sep 2026PreprintCC BY-NC-SA 4.0Full textHigh-energy & particle physicsarXiv

Cite

BibTeX RIS

BibTeX

@misc{kowalski2026machinelearned,
  title = {{Machine-learned, finite temperature Fermi-operator expansions suitable for GPUs and AI-hardware}},
  author = {Stanislaw Kowalski and Christian F. A. Negre and Anders M. N. Niklasson and Kipton Barros and Joshua Finkelstein},
  howpublished = {arXiv},
  year = {2026},
  month = sep,
  doi = {10.48550/arxiv.2605.08523},
  eprint = {2605.08523},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2605.08523},
  note = {CC0},
}

RIS

TY  - UNPB
TI  - Machine-learned, finite temperature Fermi-operator expansions suitable for GPUs and AI-hardware
AU  - Stanislaw Kowalski
AU  - Christian F. A. Negre
AU  - Anders M. N. Niklasson
AU  - Kipton Barros
AU  - Joshua Finkelstein
PY  - 2026
DA  - 2026/09/22
PB  - arXiv
DO  - 10.48550/arxiv.2605.08523
UR  - https://arxiv.org/abs/2605.08523
AB  - We present several finite-temperature recursive Fermi-operator expansion schemes based on the second-order spectral projection (SP2) method. Our approach builds on a previous observation that the electronic structure problem, as formulated through a recursive SP2 expansion, can be mapped onto the architecture of a deep neural network. Using this perspective, we generalize SP2 to finite electronic temperatures by constructing machine learning models that determine optimized recursive expansion coefficients. The same approach is also applied to the prediction of the electronic entropy for fractional occupation numbers. The coefficients are trained for a specified chemical potential and electronic temperature and are not available in closed analytical form. However, by employing an appropriate affine rescaling strategy to the Hamiltonian matrix, we eliminate the need to retrain the model for different temperatures and chemical potentials. Our approach avoids explicit diagonalization and relies solely on highly optimized matrix-matrix multiplication kernels. Compared to state-of-the-art diagonalization, we achieve an order-of-magnitude speedup in the single-particle finite-temperature density matrix calculation for small and moderately sized matrices on modern GPUs and dense matrix multiply units.
ER  -

Abstract

Related

Cite