Temporal causal chain improves reasoning in knowledge graph question answering
TCC-RAG: Temporal Causal Chain-Guided RAG for Temporal Knowledge Graph Question Answering
Computation and Language
Summary
Answering questions about events that happen over time is hard for current AI methods because they get confused by too much irrelevant information and miss important cause-and-effect steps. The authors came up with a new technique called TCC-RAG that focuses only on the key events and their causes in the right order to help find better answers. They turned complex event data into a graph that shows how events cause one another, then used this to pick the most helpful facts. This method helped AI give more accurate answers on challenging question sets about time and cause.
What this means in practice
- •For knowledge graph engineers: Improve temporal question answering by incorporating causal chains for more precise and efficient knowledge retrieval.
- •For enterprise ai developers: Enhance AI assistants that handle time-dependent queries with fewer errors by filtering retrieved knowledge based on causal consistency.
Authors
Xinxun Zhang, Yuxiang Wang, Shanshan Feng, Jun Zhang, Tieyun Qian, Binhang Yuan, Kai Yin, Xiao Yan, Jiawei Jiang
Abstract
Recent advances in large language models (LLMs) have spurred growing interest in temporal knowledge graph question answering (TKGQA). Most existing methods rely on semantic retrieval augmented generation (RAG), which suffers from fundamental limitations in complex temporal reasoning. On the one hand, semantic-based RAG methods struggle to capture the multi-hop causal knowledge required for complex temporal reasoning. On the other hand, these methods introduce excessive redundant knowledge, trapping LLMs in spurious reasoning chains that potentially result in erroneous answers. To address these problems, we propose a Temporal Causal Chain-guided RAG (TCC-RAG) to retrieve a concise yet sufficient set of causal knowledge to support complex temporal reasoning. Specifically, we reformulate the temporal knowledge graph into a multi-view causal cue graph, with causal cues inducing connections that explicitly encode temporal evolution and causal dynamics. Based on the causal cue graph, we construct a query-specific temporal causal chain to guide the retrieval process, enabling the model to focus on a concise set of causally relevant knowledge. Finally, we introduce a causal evidence filter to refine the retrieved knowledge, ensuring its causal consistency with the underlying temporal logic of the query. Experiments show that TCC-RAG achieves substantial performance gains of 10.72% and 8.40% on two TKGQA benchmarks, and delivers a remarkable 62.97% improvement for multi-step temporal reasoning.