C-Unseen: Weak Signal Detection in Dynamic Temporal Knowledge Graphs via LLM Reasoning

C-Unseen is a self‑interpretable framework designed to detect weak signals in Dynamic Temporal Knowledge Graphs (DTKGs). It defines a weak signal as a rare, semantically coherent subgraph that spreads across consecutive snapshots. The system uses a Rare Subgraphs Extractor, where a large language model identifies subgraphs that contrast with the dominant narrative through chain‑of‑thought reasoning, and a Weak Signal Alerter that tracks the persistence of these subgraphs over time to isolate true weak signals, outperforming keyword, topic, and graph‑based baselines.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.

• llms