LLM-Driven Agentic System Enhances Quantum Coherent Ising Machine
A research project combines a Coherent Ising Machine (CIM) powered by a femtosecond laser with an agentic system driven by large language models (LLMs), utilizing LangGraph and LangChain frameworks. The LLMs undertake various responsibilities, including the calibration of QUBO/Ising models, iterative decision-making for constraint weights, and swift validation of previously published methods. These tasks are executed using domestically produced large models and CIM hardware, effectively enhancing the practical application of quantum CIM in addressing NP-complete challenges.
Key facts
- Quantum computing devices are recognized as powerful tools for solving NP-complete problems.
- The study integrates a femtosecond laser-pumped Coherent Ising Machine (CIM) with an LLM-driven agentic system.
- The integration leverages LangGraph and LangChain frameworks.
- LLMs can perform QUBO/Ising model calibration, constraint weight decision iteration, and rapid validation of literature-reported schemes.
- All tasks are fully implemented based on domestic large models.
- The system combines domestic large models with domestically developed CIM hardware.
- The study aims to address barriers for non-specialists and reduce effort for experts in modeling.
- The research is published on arXiv with identifier 2605.23934.
Entities
Institutions
- arXiv