Wenbiao Tao

Wenbiao Tao, Xinyuan Li, Yunshi Lan et al.

Retrieval-Augmented Generation enhances language models by retrieving external knowledge to support informed and grounded responses. However, traditional RAG methods rely on fragment-level retrieval, limiting their ability to address query-focused summarization queries. GraphRAG introduces a graph-based paradigm for global knowledge reasoning, yet suffers from inefficiencies in information extraction, costly resource consumption, and poor adaptability to incremental updates. To overcome these limitations, we propose TagRAG, a tag-guided hierarchical knowledge graph RAG framework designed for efficient global reasoning and scalable graph maintenance. TagRAG introduces two key components: (1) Tag Knowledge Graph Construction, which extracts object tags and their relationships from documents and organizes them into hierarchical domain tag chains for structured knowledge representation, and (2) Tag-Guided Retrieval-Augmented Generation, which retrieves domain-centric tag chains to localize and synthesize relevant knowledge during inference. This design significantly adapts to smaller language models, improves retrieval granularity, and supports efficient knowledge increment. Extensive experiments on UltraDomain datasets spanning Agriculture, Computer Science, Law, and cross-domain settings demonstrate that TagRAG achieves an average winning rate of 78.36% against baselines while maintaining about 14.6x construction and 1.9x retrieval efficiency compared with GraphRAG.

Xinyuan Li, Murong Xu, Wenbiao Tao et al.

Large language models (LLMs) achieve high performance on mathematical reasoning, but these results can be inflated by training data leakage or superficial pattern matching rather than genuine reasoning. To this end, an adversarial perturbation-based evaluation is needed to measure true mathematical reasoning ability. Current rule-based perturbation methods often generate ill-posed questions and impede the systematic evaluation of question difficulty and the evolution of benchmarks. To bridge this gap, we propose RIDE, a novel adversarial question-rewriting framework that leverages Item Response Theory (IRT) to rigorously measure question difficulty and to generate intrinsically more challenging, well-posed variations of mathematical problems. We employ 35 LLMs to simulate students and build a difficulty ranker from their responses. This ranker provides a reward signal during reinforcement learning and guides a question-rewriting model to reformulate existing questions across difficulty levels. Applying RIDE to competition-level mathematical benchmarks yields perturbed versions that degrade advanced LLM performance, with experiments showing an average 21.73% drop across 26 models, thereby exposing limited robustness in mathematical reasoning and confirming the validity of our evaluation approach.

Yuanyuan Liang, Keren Tan, Tingyu Xie et al.

Graph Databases (Graph DB) find extensive application across diverse domains such as finance, social networks, and medicine. Yet, the translation of Natural Language (NL) into the Graph Query Language (GQL), referred to as NL2GQL, poses significant challenges owing to its intricate and specialized nature. Some approaches have sought to utilize Large Language Models (LLMs) to address analogous tasks like text2SQL. Nonetheless, in the realm of NL2GQL tasks tailored to a particular domain, the absence of domain-specific NL-GQL data pairs adds complexity to aligning LLMs with the graph DB. To tackle this challenge, we present a well-defined pipeline. Initially, we utilize ChatGPT to generate NL-GQL data pairs, leveraging the provided graph DB with self-instruction. Subsequently, we employ the generated data to fine-tune LLMs, ensuring alignment between LLMs and the graph DB. Moreover, we find the importance of relevant schema in efficiently generating accurate GQLs. Thus, we introduce a method to extract relevant schema as the input context. We evaluate our method using two carefully constructed datasets derived from graph DBs in the finance and medicine domains, named FinGQL and MediGQL. Experimental results reveal that our approach significantly outperforms a set of baseline methods, with improvements of 5.90 and 6.36 absolute points on EM, and 6.00 and 7.09 absolute points on EX for FinGQL and MediGQL, respectively.

Qinwen Chen, Wenbiao Tao, Zhiwei Zhu et al.

Community Question Answering (CQA) platforms can be deemed as important knowledge bases in community, but effectively leveraging historical interactions and domain knowledge in real-time remains a challenge. Existing methods often underutilize external knowledge, fail to incorporate dynamic historical QA context, or lack memory mechanisms suited for industrial deployment. We propose ComRAG, a retrieval-augmented generation framework for real-time industrial CQA that integrates static knowledge with dynamic historical QA pairs via a centroid-based memory mechanism designed for retrieval, generation, and efficient storage. Evaluated on three industrial CQA datasets, ComRAG consistently outperforms all baselines--achieving up to 25.9% improvement in vector similarity, reducing latency by 8.7% to 23.3%, and lowering chunk growth from 20.23% to 2.06% over iterations.