Query-Conditioned Graph Retrieval for Contextualized LLM Reasoning in Personalized Wearable Data

TL;DR

This paper introduces WAG, a graph-based framework that enhances LLM reasoning on wearable data by adaptively retrieving personalized, contextually relevant subgraphs, improving analysis accuracy and efficiency.

Contribution

WAG is a novel graph-based context retrieval method that organizes wearable data into personalized knowledge graphs for query-adaptive reasoning with LLMs.

Findings

WAG achieves about 70% win rate over baseline methods in real-world tests.

Structured, query-adaptive retrieval improves LLM analysis of wearable data.

Hierarchical Bayesian modeling captures both global and local relationships.

Abstract

Large language models (LLMs) are increasingly applied to analyzing wearable sensing data, which are long-term, multimodal, and highly personalized. A key challenge is context selection: providing insufficient context limits reasoning, while including all available data leads to inefficiency and degraded generation quality. We propose Wearable As Graph (WAG), a graph-based context retrieval framework that enables query-adaptive reasoning over wearable data with LLMs. WAG organizes wearable metrics and user-specific signals into a personalized knowledge graph, and retrieves a query-conditioned subgraph to support downstream generation. The retrieval process integrates global relationships, capturing prior knowledge and population- and individual-level patterns via hierarchical Bayesian modeling, with local relationships that reflect short-term signal deviations. A query openness signal further controls retrieval breadth. We evaluate WAG on over 10,000 data-grounded queries from real-world wearable datasets. Across LLM-based and human evaluations, WAG achieves an approximately 70% win rate over baseline and standard RAG methods, demonstrating the effectiveness of structured, query-adaptive context retrieval for LLM-driven analysis of wearable data.