Tag: question answering

An evaluation of 67 complete scientific abstracts found high automated scores for local question-answering checks but only 0.07 for document-wide ...

A new 200,000-question benchmark from Carnegie Mellon University reveals that leading vision-language models perform at random chance when interpreting engineering ...

Chinese researchers have built an ontology-guided LightRAG framework that converts fragmented precarious rock hazard records into a knowledge graph, achieving ...

Researchers at Xidian University have developed a Topic Semantic Relational Knowledge model that filters dialogue history by topic and matches ...

A new hybrid model called TG-Parser combines Transformer encoders with graph convolutional networks to achieve record-setting accuracy in English grammatical ...

Researchers have developed MORSE, a training-free method that treats evidence selection in retrieval-augmented question answering as a multi-objective subset optimization ...

A new neural framework called CMCE+BK combines extended candidate events with multi-level background knowledge to dramatically improve the mining of ...