en.Wedoany.com Reported - IBM Research has developed the open-source evolutionary AI framework OpenEvolve, which leverages a large language model-guided evolutionary algorithm to rapidly search for viable quantum error-correcting codes within a vast space of algebraic expressions. The search for quantum error-correcting codes is a critical bottleneck in the practical application of quantum computing, as the space of possible algebraic formulas is immense, and traditional methods are computationally intensive and resource-demanding.
The OpenEvolve framework establishes a bidirectional interaction between classical AI and quantum computing, utilizing large language models to generate hypotheses for algebraic expressions that can serve as the basis for candidate error-correcting codes. The research team tested the framework using bivariate bicycle (BB) codes, a type of quantum low-density parity-check (qLDPC) code that is a key focus area on IBM's fault-tolerant quantum computing roadmap. Through evolutionary runs, OpenEvolve discovered 465 new quantum error-correcting codes. These codes are evaluated using the [[n, k, d]] format, where n represents the number of physical qubits, k represents the number of logical qubits, and d represents the fault-tolerant distance.
Among the discovered codes, the [[288,50,8]] structure provides 50 logical qubits, breaking the previous record of up to 16 logical qubits for this code family (despite a lower distance d); the [[72,4,8]] structure requires only 72 physical qubits, making it potentially more feasible for near-term quantum hardware; and the [[288,16,12]] and [[360,12,≤24]] structures offer relatively balanced candidate solutions, competitive with IBM's extensively studied [[144,12,12]] "gross code." The research team stated that while further evaluation is needed to assess the performance of these AI-generated codes in actual physical architectures, OpenEvolve establishes a highly viable method for exploring the vast space of algebraic codes. IBM Research has fully open-sourced the framework on GitHub, encouraging the global quantum research community to utilize and expand upon it. The research has been published on the IBM Research blog and on arXiv.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
