Retrieving Time-Series Differences Using Natural Language Queries

2025-03-27

AI Summary

This paper addresses natural language (NL)-driven temporal difference retrieval—a task requiring identification of time-series segments exhibiting user-specified semantic differences (e.g., “rise-then-fall” or “increasing amplitude”) without domain-specific prior knowledge. Methodologically, we first formally define six interpretable, semantics-driven temporal difference patterns; then construct the first difference-oriented time-series–text alignment dataset; and finally propose a difference-aware multi-granularity contrastive learning framework that jointly encodes temporal dynamics and linguistic semantics. Our key contributions are threefold: (1) a human-interpretable taxonomy of temporal difference semantics; (2) the first publicly available benchmark dataset specifically designed for NL-based temporal difference retrieval; and (3) state-of-the-art performance with a mean Average Precision (mAP) of 0.994—substantially outperforming existing NL-guided time-series search methods.

Technology Category

• Natural Language Processing: Sentence-level Semantics, Textual Inference, etc.
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Machine Learning: Time-Series/Data Streams

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Effectively searching time-series data is essential for system analysis; however, traditional methods often require domain expertise to define search criteria. Recent advancements have enabled natural language-based search, but these methods struggle to handle differences between time-series data. To address this limitation, we propose a natural language query-based approach for retrieving pairs of time-series data based on differences specified in the query. Specifically, we define six key characteristics of differences, construct a corresponding dataset, and develop a contrastive learning-based model to align differences between time-series data with query texts. Experimental results demonstrate that our model achieves an overall mAP score of 0.994 in retrieving time-series pairs.

Problem

• Retrieving time-series differences via natural language queries
• Handling differences between time-series data effectively
• Aligning query texts with time-series contrastive characteristics

Innovation

• Natural language query-based time-series difference retrieval
• Six key characteristics defining time-series differences
• Contrastive learning model aligning data with queries