This repository aggregates official research projects focused on applying Artificial Intelligence to badminton analytics. It offers a suite of tools, datasets, and frameworks designed for researchers and engineers in sports science and AI, enabling detailed match analysis, player performance evaluation, and the development of intelligent agents for training and simulation. The projects aim to advance AI capabilities in understanding and predicting complex sports dynamics.

How It Works

The projects leverage deep learning models, including frameworks like ShuttleNet for stroke forecasting and DyMF for movement prediction, to capture intricate temporal and spatial dependencies within badminton rallies. It also introduces specialized reinforcement learning environments that simulate realistic game scenarios, facilitating the training of AI agents. A key component is the development and release of large-scale, stroke-level datasets, such as ShuttleSet, to support data-driven research and benchmarking.

Quick Start & Requirements

The provided README details research projects and publications but does not include explicit installation instructions, quick start commands, or specific software/hardware prerequisites. Users are likely to require standard Python environments with deep learning libraries (e.g., TensorFlow, PyTorch) and potentially GPU acceleration for many of the research implementations.

Highlighted Details

• ShuttleSet: Features the ShuttleSet, described as the largest badminton singles dataset with stroke-level records, published at KDD-23.
• Forecasting Innovations: Introduces pioneering work on stroke forecasting (AAAI-22) and movement forecasting (AAAI-23), addressing key predictive challenges in sports.
• RL Environment: Offers a dedicated CoachAI Badminton Environment (AAAI-24 Demo, DSAI4Sports @ KDD 2023) for realistic AI agent training and performance benchmarking.
• NEWSAGENT Benchmark: Presents a benchmark for evaluating LLM agents on journalistic workflows, including sports news generation (ACL Findings 2026).

Maintenance & Community

The repository is associated with research from the Advanced Database System Laboratory, National Yang Ming Chiao Tung University, supervised by Prof. Wen-Chih Peng. It lists numerous publications and contributors but does not provide links to community forums, active development roadmaps, or specific maintenance contacts.

Licensing & Compatibility

The provided README does not specify the software license for the projects or datasets. Users should exercise caution regarding usage rights, particularly for commercial applications, until licensing information is clarified.

Limitations & Caveats

This repository serves as a collection of distinct research projects rather than a unified, production-ready software package. Specific limitations for individual projects are not detailed. The NEWSAGENT project is noted with a future publication date (ACL Findings 2026), suggesting it may represent ongoing or very recent work.