Multi-hop Question Answering
A question type that requires stepwise reasoning across multiple documents, entities, and relationships to arrive at an answer. It is a representative form of complex question answering in natural language processing research.
Timeline (last 90 days)
GraphEvoResearcher proposes a training-free self-reflection protocol that lifts self-reflection meta-rewards onto an explicit reasoning graph, where nodes are intermediate conclusions/evidence and edges encode support, contradiction, and dependency relations.
Sliding windows and conservative sparse strategies easily lose cross-paragraph evidence chains, which is especially fatal in scenarios such as multi-hop question answering and legal contract analysis.
Multi-hop QA is a class of questions that require reasoning across multiple independent evidence fragments to answer, with mainstream benchmarks including HotpotQA, MuSiQue, and 2WikiMultiHopQA.
Multi-hop question answering refers to answering a question that requires step-by-step reasoning across multiple documents, multiple entities, and multiple relations.
In multi-hop scenarios, intermediate nodes are semantically distant from the original question and are not adjacent in the vector space, causing the retriever to fail to include them in the candidate set.
All Facts (5)
GraphEvoResearcher proposes a training-free self-reflection protocol that lifts self-reflection meta-rewards onto an explicit reasoning graph, where nodes are intermediate conclusions/evidence and edges encode support, contradiction, and dependency relations.
50%UnverifiedSliding windows and conservative sparse strategies easily lose cross-paragraph evidence chains, which is especially fatal in scenarios such as multi-hop question answering and legal contract analysis.
50%UnverifiedMulti-hop QA is a class of questions that require reasoning across multiple independent evidence fragments to answer, with mainstream benchmarks including HotpotQA, MuSiQue, and 2WikiMultiHopQA.
50%UnverifiedMulti-hop question answering refers to answering a question that requires step-by-step reasoning across multiple documents, multiple entities, and multiple relations.
50%UnverifiedIn multi-hop scenarios, intermediate nodes are semantically distant from the original question and are not adjacent in the vector space, causing the retriever to fail to include them in the candidate set.
50%