Oct 2026· Computers and artificial intelligence· 0 citations· 44 references
TL;DR
This review traces the evolution of CQA from retrieval and ranking through neural semantic matching and pretrained Transformer models to retrieval-augmented generation and large language models, with particular emphasis on multi-label question understanding and multi-document evidence integration.
Abstract
Community Question Answering (CQA) platforms function as continuously evolving repositories of user-generated knowledge, where questions are often informal, intents may be multiple, evidence is distributed across sources, and knowledge changes over time. These characteristics limit the applicability of conventional single-label and single-document approaches to complex information needs. This review traces the evolution of CQA from retrieval and ranking through neural semantic matching and pretrained Transformer models to retrieval-augmented generation and large language models, with particular emphasis on multi-label question understanding and multi-document evidence integration. Existing studies are compared from the perspectives of semantic representation, intent modeling, evidence retrieval and fusion, knowledge grounding, and answer reliability. The literature shows that substantial progress has been made in individual capabilities, but a unified mechanism that coordinates multi-label intent interpretation, dynamic retrieval, evidence verification, and grounded generation remains insufficiently developed. Future research should therefore move toward closed-loop architectures and evaluation frameworks that jointly measure label identification, evidence coverage, source attribution, and final answer quality in realistic community settings.
HyperDoc is proposed, a document RAG framework that constructs compact yet sufficient evidence while preserving document relations and constructs a document hypergraph to explicitly capture structural, logical, and cross-modal relationships among multimodal document elements.
Wen-Xin Zhao, Jia-Chuan Wang, Xin-Yi Zhu et al.· 0 citations
DuQUAD is proposed, a multi-document QA framework that mitigates positional bias through dual-view reasoning and quality-aware evidence pruning, and substantially reducing computational cost compared to existing multi-agent frameworks.
Yeonjun Choi, Seyeong Jeong, Jong Wook Kim et al.· Proceedings of the 2026 Conf...· 0 citations
This work proposes a TextVQA framework that integrates external prior knowledge to support multimodal reasoning and demonstrates that this approach effectively leverages external knowledge to improve performance on knowledge-intensive questions.
Existing long-document QA benchmarks rarely combine large-scale human-authored questions with detailed visual evidence annotations, limiting diagnosis of evidence localization and interpretation failures. We present VELD-QA, a large-scale benchmark for Visual Evidence grounding and question answering in Long Documents....
Zhivar Sourati, Marianne Menglin Liu, Rongguang Wang et al.· 0 citations
Document-level relation extraction must identify relations when entity mentions and supporting evidence are distributed across multiple sentences. Existing methods often use entity representations that remain fixed after encoding and provide limited support for progressively retrieving and combining dispersed evidence....
Yi Zhang, Xiu-Qi Ruan, Xin-Bin Cheng et al.· Evaluation review· 0 citations
Long-document visual question answering requires models to locate sparse question-relevant evidence across many pages and understand fine-grained textual and visual content. Processing an entire document at high resolution is computationally expensive, whereas reduced visual fidelity and irrelevant pages can impair bot...
Hong Nie, Jun Yu, Bo Wang et al.· Proceedings of the 34th ACM...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.