works on
multi-hop reasoning 1reachability estimation 1reinforcement learning 1self-supervised pretraining 1temporal knowledge graphs 1

From the 1 of 3 linked papers with an AI index.

collaborators
Showing cs.AIShow all

1 paper · 1 filter

cs.AI2026

Reachability-Aware Pretraining for Efficient Target-Oriented Path Exploration in Temporal Knowledge Graph Reasoning

Chien-Liang Liu, Tsao-Lun Chen

The paper introduces RAPTOR, a self‑supervised pretraining technique that estimates the reachability of actions to a target entity, improving exploration efficiency and performance…