The Retrieval-Augmented Generation (RAG) framework enhances Large Language Models (LLMs) by retrieving relevant knowledge to broaden their knowledge boundaries and mitigate factual hallucinations stemming from knowledge gaps. However, the RAG Framework faces challenges in effective knowledge retrieval and utilization; invalid or misused knowledge will interfere with LLM generation, reducing reasoning efficiency and answer quality. Existing RAG methods address these issues by decomposing and expanding queries, introducing special knowledge structures, and using reasoning process evaluation and feedback. However, the linear reasoning structures limit complex thought transformations and reasoning based on intricate queries. Additionally, knowledge retrieval and utilization are decoupled from reasoning and answer generation, hindering effective knowledge support during answer generation. To address these limitations, we propose the CRP-RAG framework, which employs reasoning graphs to model complex query reasoning processes more comprehensively and accurately. CRP-RAG guides knowledge retrieval, aggregation, and evaluation through reasoning graphs, dynamically adjusting the reasoning path based on evaluation results and selecting knowledge-sufficiency paths for answer generation. CRP-RAG outperforms the best LLM and RAG baselines by 2.46 in open-domain QA, 7.43 in multi-hop reasoning, and 4.2 in factual verification. Experiments also show the superior factual consistency and robustness of CRP-RAG over existing RAG methods. Extensive analyses confirm its accurate and fact-faithful reasoning and answer generation for complex queries.

The Retrieval-Augmented Generation (RAG) Framework enhances Large Language Models (LLMs) by retrieving relevant knowledge to broaden their knowledge boundaries and mitigate factual hallucinations stemming from knowledge gaps. However, the RAG Framework faces challenges in effective knowledge retrieval and utilization; invalid or misused knowledge will interfere with LLM generation, reducing reasoning efficiency and answer quality. Existing RAG methods address these issues by decomposing and expanding queries, introducing special knowledge structures, and using reasoning process evaluation and feedback. However, the linear reasoning structures limit complex thought transformations and reasoning based on intricate queries. Additionally, knowledge retrieval and utilization are decoupled from reasoning and answer generation, hindering effective knowledge support during answer generation. To address these limitations, we propose the CRP-RAG framework, which employs reasoning graphs to model complex query reasoning processes more comprehensively and accurately. CRP-RAG guides knowledge retrieval, aggregation, and evaluation through reasoning graphs, dynamically adjusting the reasoning path based on evaluation results and selecting knowledge-sufficiency paths for answer generation. Experimental results show that CRP-RAG significantly outperforms state-of-the-art LLMs and RAG baselines in three reasoning and question answering tasks, demonstrating superior factual consistency and robustness compared to existing RAG methods.

Uncrewed aerial vehicles (UAVs) ad-hoc network is expected to play an important role in the integrated sensing and communication (ISAC) due to its flexible topology and low cost, enabling scalable connectivity. However, the Line of Sight and node mobility of UAVs lead to severe wireless interference and highly dynamic network topology, making the design of distributed medium access control protocols for large-scale UAV ad-hoc network a critical problem. Here, we propose H-SATMAC, a hybrid self-adaptive time-division multiple access-based MAC protocol for large-scale UAV ad-hoc networks. In order to jointly consider the transmission of the periodic and aperiodic packets, we design a hybrid channel resource scheduling method based on time slot groups. The consecutive free time slots in the group are used as the contention window for CSMA-based transmissions. We utilize a dynamic mapping of slot groups and the geohash coding system to avoid hidden terminal problem and load-balance multiple group resources. Besides, an enhanced time slot adjustment and adaptive frame length method is designed to achieve better channel utilization while avoiding channel collisions between nodes. We conduct theoretical analysis and extensive simulations to evaluate H-SATMAC. Simulation results show that H-SATMAC achieves 22% higher channel utilization and above 64% larger number of total reception packets than that of the reference protocol.

Despite the remarkable progress of large language models (LLMs) in understanding and generating unstructured text, their application in structured data domains and their multi-role capabilities remain underexplored. In particular, utilizing LLMs to perform complex reasoning tasks on knowledge graphs (KGs) is still an emerging area with limited research. To address this gap, we propose KG-EGV, a versatile framework leveraging LLMs to perform KG-based tasks. KG-EGV consists of four core steps: sentence segmentation, graph retrieval, EGV, and backward updating, each designed to segment sentences, retrieve relevant KG components, and derive logical conclusions. EGV, a novel integrated framework for LLM inference, enables comprehensive reasoning beyond retrieval by synthesizing diverse evidence, which is often unattainable via retrieval alone due to noise or hallucinations. The framework incorporates six key stages: generation expansion, expansion evaluation, document re-ranking, re-ranking evaluation, answer generation, and answer verification. Within this framework, LLMs take on various roles, such as generator, re-ranker, evaluator, and verifier, collaboratively enhancing answer precision and logical coherence. By combining the strengths of retrieval-based and generation-based evidence, KG-EGV achieves greater flexibility and accuracy in evidence gathering and answer formulation. Extensive experiments on widely used benchmarks, including FactKG, MetaQA, NQ, WebQ, and TriviaQA, demonstrate that KG-EGV achieves state-of-the-art performance in answer accuracy and evidence quality, showcasing its potential to advance QA research and applications.

Given the remarkable text generation capabilities of pre-trained language models, impressive results have been realized in graph-to-text generation. Howeve