Quantum circuits as a game: A reinforcement learning agent for quantum compilation arXiv:2506.05536

Part of the "Science with QuEra" webinar series, co-authors from NVIDIA and QuEra discuss the latest paper: arXiv:2506.05536

Join us in this enlightening episode of "Science with QuEra" as Yuval Boger of QuEra, hosts a groundbreaking discussion featuring Jonathan Wurtz from QuEra and Kohei Nakaji from NVIDIA. Discover how the collaboration between QuEra and NVIDIA is setting new standards in quantum computing with their innovative project: "Quantum Circuits as a Game: A Reinforcement Learning Agent for Quantum Compilation."

In this webinar, we start by exploring QuEra's cutting-edge quantum computing technology based on neutral atom arrays and how these systems promise scalability and high fidelity. Learn about the unique advantages of neutral atoms, such as their perfect uniformity and room-temperature operation, making them a promising path toward fault-tolerant quantum computers.

The highlight of the session is the introduction of the "Atom Game," a revolutionary approach employing reinforcement learning to optimize quantum circuit compilation. Watch as our experts explain the intricacies of this game, where an AI agent, the Quantum Circuit Demon, learns to efficiently configure atom layouts, optimizing for future quantum gate executions.

Key topics include:

- The benefits of neutral atom technology in quantum computing.
- The concept of mid-circuit reconfigurability and its impact on scalability.
- How reinforcement learning is employed to improve quantum circuit compilation.
- Insights into the collaboration between QuEra and NVIDIA, including their joint efforts in the NVIDIA Accelerated Quantum Center (NVAQC).

Chappters:
00:00 - Intro to Science with QuEra
00:26 - Overview of QuEra and Neutral Atom Computing
02:44 - Why Neutral Atoms? Key Advantages
04:16 - Mid-Circuit Reconfigurability and Gemini Device
07:41 - Logical Magic State Distillation Demonstration
09:20 - AOD Constraints and Conflict Graphs
13:49 - Reconfiguration Costs and Move Synthesis
15:27 - Why Use Reinforcement Learning?
18:13 - NVIDIA's Role in Quantum + AI
19:09 - Introducing the Atom Game
21:02 - AI Model Objectives and Constraints
24:29 - Model Architecture: Transformer-Based Design

#QuantumComputing #ReinforcementLearning #NeutralAtoms
28:12 - Experiment 1: Training on Single Circuits
31:33 - Experiment 2: Generalization to New Circuits
34:08 - Open Questions and Future Directions
36:11 - Summary and Collaborations with NVIDIA
37:26 - Closing Remarks

Видео Quantum circuits as a game: A reinforcement learning agent for quantum compilation arXiv:2506.05536 канала QuEra Computing