Quantum annealing-assisted framework for Search-LWE: a proof-of-concept study on a real quantum computer
The Learning With Errors (LWE) problem is fundamental in post-quantum cryptography, and its hardness ensures the security of many lattice-based cryptosystems. This paper proposes a quantum-classical framework for solving Search-LWE instances on real quantum annealing hardware using quantum tunneling to explore a restricted correction space around the Babai solution. First, the Search-LWE is formulated as a lattice decoding problem, and the restricted binary correction space around the Babai solution is encoded into an Ising Hamiltonian. Then, forward and reverse quantum annealing are implemented on the D-Wave Advantage2 to explore the correction space. In particular, reverse quantum annealing uses the Babai solution as the initial state for quantum annealing. The outputs of both quantum annealing schemes are refined by a classical local search to improve solution quality. Experimental results show that both quantum annealing approaches obtain lattice candidates with smaller residual error norms than the Babai solution for a consider-able number of tested instances. Simulated annealing further confirms the effectiveness of the Ising formulation. Moreover, our framework demonstrates successful secret recovery for Search-LWE instances with dimensions up to 15 on D-Wave Advantage2. These results provide a proof of concept on quantum hardware, but do not demonstrate a practical threat to cryptographically relevant LWE-based cryptosystems.
Authors
- Chunlei Hong
- Shuaishuai Zhu (ORCID: https://orcid.org/0000-0003-1573-3808)
- Han Yiliang
- Yibo Li
- Shuaizhi Xiong
- Zhi Pei
- Yidan Chen
- Chao Wang
Institutions
- Shanghai University (CN)
Publication Details
- Journal
- Tsinghua Science & Technology
- Published
- 2026-11-18
- DOI
- https://doi.org/10.26599/tst.2026.9010101
- Primary Topic
- Quantum Computing Algorithms and Architecture
- Type
- article
- Field-Weighted Citation Impact
- 0.00