QUANTA: Quantum Network AgenTic Arena
Abstract
The advent of modern AI and large language models (LLMs) applied to scientific research and recent progress in domains such as Physics, Biology, and Mathematics, has made agentic science an area of increasing interest. In this paper, we introduce the Quantum Network AgenTic Arena (QUANTA), an agentic system designed to support quantum networking research. QUANTA integrates the SeQUeNCe quantum network simulator, an execution environment, literature APIs, and domain knowledge resources-derived from prior research publications within our group-to enable simulation-driven investigation. The system supports literature analysis, experiment design, simulation execution, and evaluation by providing agents with the necessary tools and resources. We introduce the Planner & Executor workflow as an assessment framework for guiding the design of the primary Discovery workflow, a five-agent architecture inspired by the scientific method for quantum networking research. We evaluate the Discovery workflow through a preliminary assessment of its operational behavior, using the findings to identify improvements for a full-scale evaluation. QUANTA provides an initial framework for studying how agentic AI can support research in this domain through the integration of simulation tools and specialized knowledge resources. To the best of our knowledge, this work is the first to investigate the application of agentic AI to quantum networking research. Evaluation of research objectives using QUANTA is ongoing and is reserved for future work.