Foresight-over-Graph: Reasoning Beyond Local Horizons for Knowledge Base Question Answering
To address this limitation, we propose Foresight-over-Graph (FoG), a foresight-aware evidence retrieval framework for knowledge base question answering (KBQA).
ProofPaper ↗
Key points
- Large language models (LLMs) have demonstrated strong capabilities in question answering, yet they still frequently suffer from hallucinations on knowledge-intensive tasks.
- Knowledge graphs (KGs) provide LLMs with structured, interpretable, and updatable factual grounding, making them a promising external knowledge source for reliable reasoning.
- However, existing LLM-guided graph reasoning methods typically rely on hop-wise greedy or beam-style pruning during evidence retrieval.
- FoG iteratively constructs a question-relevant evidence subgraph and uses far-to-near feedback to guide path exploration, and maintains a compact memory subgraph to support continued exploration.
Sources (1)
- [1]Foresight-over-Graph: Reasoning Beyond Local Horizons for Knowledge Base Question AnsweringarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 6, 02:07 PM
To address this limitation, we propose Foresight-over-Graph (FoG), a foresight-aware evidence retrieval framework for knowledge base question answering (KBQA).
Large language models (LLMs) have demonstrated strong capabilities in question answering, yet they still frequently suffer from hallucinations on knowledge-intensive tasks.
Extractive summary: sentences quoted from the sources.