Zhao Tan, Yiji Zhao, Shiyu Wang, Chang Xu, Yuxuan Liang, Xiping Liu, Shirui Pan, Ming Jin
Sonar-TS is a neuro-symbolic framework for natural language querying of time series databases that uses a Search-Then-Verify approach to handle complex temporal queries effectively.
Sonar-TS is a new system designed to help people who aren't experts in databases to easily search and analyze time-based data using natural language. This is especially useful for finding patterns or unusual events in large sets of time series data, like detecting anomalies in financial records or monitoring environmental changes. The system works by first searching for potential matches using SQL, then verifying these matches with Python programs to ensure accuracy. Additionally, the creators have developed a new benchmark to evaluate how well such systems perform, which will aid future research in this area.
