Large Language Model-Assisted Superconducting Qubit Experiments
Authors: Shiheng Li, Jacob M. Miller, Phoebe J. Lee, Gustav Andersson, Christopher R. Conner, Yash J. Joshi, Bayan Karimi, Amber M. King, Howard L. Malc, Harsh Mishra, Hong Qiao, Minseok Ryu, Xuntao Wu, Siyuan Xing, Haoxiong Yan, Jian Shi, Andrew N. Cleland
Organizations: Department of Physics, University of Chicago, Chicago, IL 60637, USA · Pritzker School of Molecular Engineering, University of Chicago, Chicago, IL 60637, USA · Pico group, QTF Centre of Excellence, Department of Applied Physics, Aalto University School of Science, P.O. Box 13500, 00076 Aalto, Finland · Department of Mathematics, University of Chicago, Chicago, IL 60637, USA · Department of Materials Science and Engineering, Rensselaer Polytechnic Institute, Troy, NY 12180, USA
Superconducting circuits have demonstrated significant potential in quantum information processing and quantum sensing. Implementing novel control and measurement sequences for superconducting qubits is often a complex and time-consuming process, requiring extensive expertise in both the underlying physics and the specific hardware and software. In this work, we introduce a framework that leverages a large language model (LLM) to automate qubit control and measurement. Specifically, our framework conducts experiments by generating and invoking schema-less tools on demand via a knowledge base on instrumental usage and experimental procedures. We showcase this framework with two experiments: an autonomous resonator characterization and a direct reproduction of a quantum non-demolition (QND) characterization of a superconducting qubit from literature. This framework enables rapid deployment of standard control-and-measurement protocols and facilitates implementation of novel experimental procedures, offering a more flexible and user-friendly paradigm for controlling complex quantum hardware.
Figures & tables
Figure 1: Schematic overview of the experimental setup. The measurement system is organized into a multi-layer architecture that bridges high-level user intent with physical quantum circuits. Via the experiment computer, both users and AI systems can interact with specialized control packages, most notably the open-source package QuICK [ 25 ] . The experiment computer is connected to a local area network (LAN), through which commands are transmitted to a suite of instruments, which synthesize and detect analog RF and DC signals routed to a dilution refrigerator (DR) to interact with superconducting circuits at cryogenic temperatures. Experimental progress and results are managed by a centralized storage server and visualized in real-time through the open-source Grapher software [ 9 ] . In practice, multiple sets of computers, instruments, and DRs are connected to the same LAN to allow convenient and simultaneous experiments across the laboratory.
Figure 2: Architecture of Heuristic Autonomous Lab (HAL). (a) Workflow: The system operates in a cycle with multiple steps, supported by long-term (green) and short-term (amber) contexts. High-thinking steps (purple) involve complex tasks including planning and development, while low-thinking steps (blue) handle routine tasks like searching. (b) Search Agent: Iterative retrieval-augmented generation (RAG) process to accurately find relevant documents from the knowledge base using vectorization and cosine similarity, effectively handling references among documents by iteration. (c) Execution Runtime: A sandbox environment, where the generated code is invoked, maintains a persistent state across executions and allows the invocation of other codes. (d) Signal Pathway: A feedback mechanism to close the information loop, allowing the AI system to acquire critical information about the execution results.
Figure 3: Autonomous resonator characterization. (a) Summary of the user input (blue), performed actions (purple), STATE read and write (black), and resulting signals (red) across five cycles. Cycles 2, 4, and 5 run without any user input. (b) Measured spectrum corresponding to the vector network analyzer (VNA) scan in cycle 1 (orange) and cycle 3 (blue). Pink spans are the resonators identified after cycle 5. Inset shows a representative circle fit performed during cycle 5 [ 19 ] .
Figure 4: Implementation of a QND characterization from a published journal article [ 12 ] . (a) Workflow for translating scientific literature into a working experiment. A journal article is processed by an LLM chatbot to generate lab-independent instructions, which are then refined into lab-specific instructions by the HAL answer component, guiding HAL on how to conduct the experiment. (b) Experimental results of a QND characterization sequence. The main plot shows the average correlation as a function of the readout cycle index. Data (blue) is fitted against the decay model [ 12 ] , where a leakage rate from the computational basis of the qubit L=0.124±0.017 is extracted from the fit (red). Inset illustrates the pulse sequence (not to scale), showing the randomized qubit control π pulses (amber) and readout pulses (green).
Figure 5: Readout fidelity metrics [ 28 , 31 , 12 ] (visibility, repeatability, and complement of leakage ( 1−L )) as a function of readout power, the latter in units of the QICK board channel gain. The solid lines are polylines for visual guidance. The plot shows improvement of visibility and repeatability with readout power until near the maximum shown here, but general degradation due to increased leakage as power is increased.
Large language models have recently shown potential in bridging the gap between classical machine learning and quantum machine learning. However, the lack of standardized, high-quality datasets and robust translation frameworks limits progress in this domain. We introduce Q-Bridge, an LLM-guided code translation framework that systematically converts CML implementations into executable QML variants. Our approach builds on a self-involving pipeline that iteratively expands a verified seed codebase into a large-scale dataset, CML-2-QML, integrating verifiable and unverifiable code pairs. The Q-Bridge model is fine-tuned using supervised LoRA adaptation for scalable and memory-efficient training, achieving faithful and interpretable quantum code generation across diverse architectures. Empirical analysis confirms the feasibility of direct CML-to-QML translation and reveals consistent structural alignment between classical and quantum paradigms. Case studies further demonstrate that Q-Bridge can maintain deterministic correctness and also enable creative architectural exploration. This work establishes the first reproducible framework and dataset for LLM-driven quantum code translation, offering a foundation for scalable quantum AI development.
Quantum computing devices are recognized as powerful tools for solving NP-complete problems. However, the intricacy of their modeling presents notable barriers for non-specialists, while the tedious iteration of constraint weights and modeling methodologies also consumes substantial effort on the part of experts. To address these challenges, this study integrates a femtosecond laser-pumped Coherent Ising Machine (CIM) with an LLM-driven agentic system by leveraging the LangGraph and LangChain frameworks. Comprehensive investigations demonstrate that large language models (LLMs) can effectively perform such tasks in modeling as QUBO/Ising model calibration, constraint weight decision iteration and rapid validation of literature-reported schemes. Notably, all these tasks can be fully implemented based on domestic large models, combined with domestically developed CIM hardware, we truly achieve the practical empowerment of quantum CIM that fully relies on all-domestic agentic large models and hardware. This work successfully realizes robust technological integration, laying a solid foundation for subsequent research. Nevertheless, it also identifies the persisting challenges in the two cutting-edge fields of large models and quantum computing at the current stage. Encouragingly, we unexpectedly discover a promising new paradigm where accumulated knowledge from agent-assisted quantum computing iterations reciprocally enhances the agent's own problem-solving capability, thereby addressing these challenges.
Wang Rui, Lu Diannan
Department of Chemical Engineering, Tsinghua University, Beijing, 100084, China.
We introduce HAML (Hamiltonian Adaptation via Meta-Learning), a framework for fast online adaptation of effective Hamiltonian models of superconducting quantum processors. HAML proceeds in two phases. A supervised training phase uses an ensemble of simulated devices to learn an offline map from control inputs and device parameters to effective Hamiltonian coefficients. An online adaptation phase then uses a small number of hardware-accessible measurements to identify the unknown parameters of a new device. By training directly against effective two-qubit coefficients extracted from full multi-mode simulations, HAML implicitly learns the reduction from full multi-mode Hamiltonians to effective qubit descriptions without invoking perturbation theory. We further show that a variance-maximizing greedy selection of measurement configurations boosts online adaptation efficiency. We demonstrate HAML on a transmon-coupler-transmon system, recovering effective two-qubit coefficients across a wide range of operating regimes, including parameter regions where Schrieffer-Wolff perturbation theory (SWPT) breaks down. This establishes a scalable, sample-efficient approach to Hamiltonian reduction and characterization for near-term quantum processors, with direct implications for calibration, control, and error mitigation.
Arielle Sanford, Andrew T. Kamen, Frederic T. Chong +1
Department of Computer Science, University of Chicago, Chicago, IL 60637 · Pritzker School of Molecular Engineering, University of Chicago, Chicago, IL 60637 · Johns Hopkins Applied Physics Laboratory, Laurel, MD 20723