QuICS Seminar: Jiaqi Leng
Description
Title:Â Quantum score matching with applications to learning thermal states
Speaker:  Jiaqi Leng (Virginia Tech)
Date & Time:Â Â October 28, 2026, 11:00am
Where to Attend:Â Â ATL 3100A and Virtual Via Zoom: https://umd.zoom.us/j/98531138611
Score matching has driven major advances in classical generative learning by enabling models to learn from data without evaluating intractable partition functions. Extending this principle to quantum learning requires to handle non-commutative density operators, which creates fundamental challenges in both defining quantum scores and developing a training framework with efficient circuit implementations and rigorous theoretical guarantees. In this work, we bridge this gap by establishing a general quantum score-matching framework with end-to-end guarantees. Applied to thermal-state learning, our approach avoids additional quantum Gibbs sampling and achieves information-theoretically optimal sample complexity in the high-temperature regime for Hamiltonians with bounded locality and interaction degree. Furthermore, numerical simulations and demonstration on the IBM quantum computer demonstrate the effectiveness and robustness of this method with limited measurement budgets and mild hardware noise. These results extend score matching into an experimentally realizable paradigm for quantum state learning. This talk is based on arXiv:2609.28391.
*We strongly encourage attendees to use their full name (and if possible, their UMD credentials) to join the zoom session.*
Speaker:  Jiaqi Leng (Virginia Tech)
Date & Time:Â Â October 28, 2026, 11:00am
Where to Attend:Â Â ATL 3100A and Virtual Via Zoom: https://umd.zoom.us/j/98531138611
Score matching has driven major advances in classical generative learning by enabling models to learn from data without evaluating intractable partition functions. Extending this principle to quantum learning requires to handle non-commutative density operators, which creates fundamental challenges in both defining quantum scores and developing a training framework with efficient circuit implementations and rigorous theoretical guarantees. In this work, we bridge this gap by establishing a general quantum score-matching framework with end-to-end guarantees. Applied to thermal-state learning, our approach avoids additional quantum Gibbs sampling and achieves information-theoretically optimal sample complexity in the high-temperature regime for Hamiltonians with bounded locality and interaction degree. Furthermore, numerical simulations and demonstration on the IBM quantum computer demonstrate the effectiveness and robustness of this method with limited measurement budgets and mild hardware noise. These results extend score matching into an experimentally realizable paradigm for quantum state learning. This talk is based on arXiv:2609.28391.
*We strongly encourage attendees to use their full name (and if possible, their UMD credentials) to join the zoom session.*