Learning functions of Hamiltonians with Hamiltonian Fourier features
We propose a quantum machine learning task that is provably easy for quantum computers and hard for classical ones. The task involves predicting quantities of the form T r [ f ( H ) ρ ] , where f is an unknown function, given descriptions of H and ρ . Using a Fourier-based feature map of Hamiltonians and linear regression, we theoretically establish the learnability of the task and implement it on a superconducting device using up to 72 qubits. This work provides a machine learning task with practical relevance, provable quantum easiness, and near-term feasibility.
Authors
- Kosuke Mitarai (ORCID: https://orcid.org/0000-0002-9056-316X)
- Hidetaka Manabe (ORCID: https://orcid.org/0009-0002-2302-0357)
- Akimoto Nakayama (ORCID: https://orcid.org/0009-0004-9684-1804)
- Yuto Morohoshi
Publication Details
- Journal
- Physical Review Letters
- Published
- 2026-11-18
- DOI
- https://doi.org/10.1103/4p8x-lk4r
- Citations
- 3
- Primary Topic
- Quantum Computing Algorithms and Architecture
- Type
- article
- Field-Weighted Citation Impact
- 1.94
Funders
- Ministry of Education, Culture, Sports, Science and Technology
- Japan Society for the Promotion of Science
- Core Research for Evolutional Science and Technology
- Co-creation place formation support program