The integration of knowledge graphs (KGs) with large language models (LLMs) offers significant potential to enhance the retrieval stage in retrieval-augmented generation (RAG) systems. In this study, we propose KG-CQR, a novel framework for Contextual Query Retrieval (CQR) that enhances the retrieval phase by enriching complex input queries with contextual representations derived from a corpus-centric KG. Unlike existing methods that primarily address corpus-level context loss, KG-CQR focuses on query enrichment through structured relation representations, extracting and completing relevant KG subgraphs to generate semantically rich query contexts. Comprising subgraph extraction, completion, and contextual generation modules, KG-CQR operates as a model-agnostic pipeline, ensuring scalability across LLMs of varying sizes without additional training. Experimental results on the RAGBench and MultiHop-RAG datasets demonstrate that KG-CQR outperforms strong baselines, achieving improvements of up to 4–6% in mAP and approximately 2–3% in Recall@25. Furthermore, evaluations on challenging RAG tasks such as multi-hop question answering show that, by incorporating KG-CQR, the performance outperforms the existing baseline in terms of retrieval effectiveness.

Integrating knowledge graphs (KGs) to enhance the reasoning capabilities of large language models (LLMs) is an emerging research challenge in claim verification. While KGs provide structured, semantically rich representations well-suited for reasoning, most existing verification methods rely on unstructured text corpora, limiting their ability to effectively leverage KGs. Additionally, despite possessing strong reasoning abilities, modern LLMs struggle with multi-step modular pipelines and reasoning over KGs without adaptation. To address these challenges, we propose ClaimPKG, an end-to-end framework that seamlessly integrates LLM reasoning with structured knowledge from KGs. Specifically, the main idea of ClaimPKG is to employ a lightweight, specialized LLM to represent the input claim as pseudo-subgraphs, guiding a dedicated subgraph retrieval module to identify relevant KG subgraphs. These retrieved subgraphs are then processed by a general-purpose LLM to produce the final verdict and justification. Extensive experiments on the FactKG dataset demonstrate that ClaimPKG achieves state-of-the-art performance, outperforming strong baselines in this research field by 9%-12% accuracy points across multiple categories. Furthermore, ClaimPKG exhibits zero-shot generalizability to unstructured datasets such as HoVer and FEVEROUS, effectively combining structured knowledge from KGs with LLM reasoning across various LLM backbones.

Claim verification is a long-standing and challenging task that demands not only high accuracy but also explainability and thoroughness of the verification process. This task becomes an emerging research issue in the era of large language models (LLMs) since real-world claims are often complex, featuring intricate semantic structures or obfuscated entities. Traditional approaches typically address this by decomposing claims into sub-claims and querying a knowledge base to resolve hidden or ambiguous entities. However, the absence of effective disambiguation strategies for these entities can compromise the entire verification process. To address these challenges, we propose Verify-in-the-Graph (VeGraph), a novel framework leveraging the reasoning and comprehension abilities of LLM agents. VeGraph operates in three phases: (1) Graph Representation - an input claim is decomposed into structured triplets, forming a graph-based representation that integrates both structured and unstructured information; (2) Entity Disambiguation -VeGraph iteratively interacts with the knowledge base to resolve ambiguous entities within the graph for deeper sub-claim verification; and (3) Verification - remaining triplets are verified to complete the fact-checking process. Experiments using Meta-Llama-3-70B (instruct version) show that VeGraph achieves competitive performance compared to baselines across benchmarks (HoVer and FEVEROUS), effectively addressing claim verification challenges. Our source code and data are available for further exploitation.

Task-oriented dialogue (TOD) systems facilitate goal-driven interactions between users and machines. While recent advances in deep learning have improved the performance, TOD systems often struggle in low-resource scenarios with limited labeled data. To address this challenge, we propose Spec-TOD, a novel framework designed to train an end-to-end TOD system with limited data. Spec-TOD introduces two main innovations: (i) a novel specialized end-to-end TOD framework that incorporates explicit task instructions for instruction-tuned large language models (LLMs), and (ii) an efficient training strategy that leverages lightweight, specialized LLMs to achieve strong performance with minimal supervision. Experiments on the MultiWOZ dataset, a widely used TOD benchmark, demonstrate that Spec-TOD achieves competitive results while significantly reducing the need for labeled data. These findings highlight the potential of the proposed framework in advancing efficient and effective TOD systems in low-resource settings.

This study introduces Simple Reasoning with Code (SiRC), a novel instruction fine-tuning method for solving mathematical reasoning problems, particularly effective for Vietnamese, which is considered a low-resource language. Specifically, solving mathematical problems requires strategic and logical reasoning, which remains challenging in this research area. This paper presents a simple yet effective instruction fine-tuning method for mathematical reasoning. Unlike previous approaches, our proposed method effectively combines chain-of-thought reasoning with code transfer methods without requiring a sophisticated inference procedure. Furthermore, we focus on exploiting small open-source large language models (LLMs) for the Vietnamese language. In this regard, we first introduce a trainable Vietnamese math reasoning dataset, which is named ViMath-InstructCode. The proposed dataset is then used for fine-tuning open-source LLMs (e.g., less than 10 billion parameters). Experiments conducted on our custom ViMath-Bench dataset, the largest benchmarking dataset focusing on Vietnamese mathematical problems, indicate the promising results of our proposed method. Our source code and dataset are available for further exploitation.

Task-oriented dialogue (TOD) systems are introduced to solve specific tasks, which focus on training multiple tasks such as language understanding, tracking states, and generating appropriate responses to help users achieve their specific goals. Currently, one of the remaining challenges in this emergent research field is the capability to produce more robust architectures fine-tuned for end-to-end TOD systems. In this study, we consider this issue by exploiting the ability of pre-trained models to provide synthesis responses, which are then used as the input for the fine-tuned process. The main idea is to overcome the gap between the training process and inference process during fine-tuning end-to-end TOD systems. The experiment on Multiwoz datasets shows the effectiveness of our model compared with strong baselines in this research field. The source code is available for further exploitation.

Court Judgement Prediction with Explanation (CJPE) is a task in the field of legal analysis and evaluation, which involves predicting the outcome of a court case based on the available legal text and providing a detailed explanation of the prediction. This is an important task in the legal system as it can aid in decision-making and improve the efficiency of the court process. In this paper, we present a new approach to understanding legal texts, which are normally long documents, based on data-oriented methods. Specifically, we first try to exploit the characteristic of data to understand the legal texts. The output is then used to train the model using the Longformer architecture. Regarding the experiment, the proposed method is evaluated on the sub-task CJPE of the SemEval-2023 Task 6. Accordingly, our method achieves top 1 and top 2 on the classification task and explanation task, respectively. Furthermore, we present several open research issues for further investigations in order to improve the performance in this research field.

Rhetorical Roles (RR) prediction is to predict the label of each sentence in legal documents, which is regarded as an emergent task for legal document understanding. In this study, we present a novel method for the RR task by exploiting the long context representation. Specifically, legal documents are known as long texts, in which previous works have no ability to consider the inherent dependencies among sentences. In this paper, we propose GNNRR (Graph Neural Network for Rhetorical Roles Prediction), which is able to model the cross-information for long texts. Furthermore, we develop multitask learning by incorporating label shift prediction (LSP) for segmenting a legal document. The proposed model is evaluated on the SemEval 2023 Task 6 - Legal Eval Understanding Legal Texts for RR sub-task. Accordingly, our method achieves the top 4 in the public leaderboard of the sub-task. Our source code is available for further investigation https://github.com/hiepnh137/SemEval2023-Task6-Rhetorical-Roles.

Heterogeneous Graph Neural Networks (HeterGNN) have been recently introduced as an emergent approach for extracting document summarization (EDS) by exploiting the cross-relations between words and sentences. However, applying HeterGNN for long documents is still an open research issue. One of the main majors is the lacking of inter-sentence connections. In this regard, this paper exploits how to apply HeterGNN for long documents by building a graph on sentence-level nodes (homogeneous graph) and combine with HeterGNN for capturing the semantic information in terms of both inter and intra-sentence connections. Experiments on two benchmark datasets of long documents such as PubMed and ArXiv show that our method is able to achieve state-of-the-art results in this research field.

Graph Neural Network (GNN)-based models have proven effective in various Natural Language Processing (NLP) tasks in recent years. Specifically, in the case of the Extractive Document Summarization (EDS) task, modeling documents under graph structure is able to analyze the complex relations between semantic units (e.g., word-to-word, word-to-sentence, sentence-to-sentence) and enrich sentence representations via valuable information from their neighbors. However, long-form document summarization using graph-based methods is still an open research issue. The main challenge is to represent long documents in a graph structure in an effective way. In this regard, this paper proposes a new heterogeneous graph neural network (HeterGNN) model to improve the performance of long document summarization (HeterGraphLongSum). Specifically, the main idea is to add the passage nodes into the heterogeneous graph structure of word and sentence nodes for enriching the final representation of sentences. In this regard, HeterGraphLongSum is designed with three types of semantic units such as word, sentence, and passage. Experiments on two benchmark datasets for long documents such as Pubmed and Arxiv indicate promising results of the proposed model for the extractive long document summarization problem. Especially, HeterGraphLongSum is able to achieve state-of-the-art performance without relying on any pre-trained language models (e.g., BERT). The source code is available for further exploitation on the Github.