Feiyang Li

Feiyang Li, Yile Wang

Task vectors, representing directions in model or activation spaces that encode task-specific behaviors, have emerged as a promising tool for steering large language models (LLMs). However, existing approaches typically require fine-tuning or invasive manipulation of internal states, limiting their flexibility and scalability. We propose \textsc{DeCoVec} (Decoding Space based Task Vector), a training-free and non-invasive framework that constructs task vectors directly in the \textit{decoding space} by leveraging in-context learning (ICL). Specifically, \textsc{DeCoVec} captures the task essence as the difference between the output logit distributions of few-shot and zero-shot prompts, then steers generation by injecting this vector into the decoding process. Experiments across seven LLMs (0.5B--9B) on TruthfulQA, Math-500, and AQUA-RAT show that \textsc{DeCoVec} consistently outperforms standard few-shot baselines, with gains up to +5.50 average accuracy. Further analysis demonstrates that \textsc{DeCoVec} effectively suppresses generation degeneration and logical flaws while exhibiting strong robustness to demonstration ordering, all without incurring additional input token costs. Our method offers a training-free and non-invasive solution for LLM steering without requiring weight updates or auxiliary models.

Feiyang Li, Peng Fang, Zhan Shi et al.

Chain-of-thought (CoT) reasoning boosts large language models' (LLMs) performance on complex tasks but faces two key limitations: a lack of reliability when solely relying on LLM-generated reasoning chains and lower reasoning performance from natural language prompts compared with code prompts. To address these issues, we propose CoT-RAG, a novel reasoning framework with three key designs: (i) Knowledge Graph-driven CoT Generation, featuring knowledge graphs to modulate reasoning chain generation of LLMs, thereby enhancing reasoning credibility; (ii) Learnable Knowledge Case-aware RAG, which incorporates retrieval-augmented generation (RAG) into knowledge graphs to retrieve relevant sub-cases and sub-descriptions, providing LLMs with learnable information; (iii) Pseudo Program Prompting Execution, which promotes greater logical rigor by guiding LLMs to execute reasoning tasks as pseudo-programs. Evaluations on nine public datasets spanning three reasoning tasks reveal significant accuracy gains-ranging from 4.0% to 44.3%-over state-of-the-art methods. Furthermore, tests on four domain-specific datasets demonstrate exceptional accuracy and efficient execution, underscoring its practical applicability and scalability. Our code and data are available at https: //github.com/hustlfy123/CoT-RAG.

Yixiao Peng, Hao Hu, Feiyang Li et al.

While virtualization and resource pooling empower cloud networks with structural flexibility and elastic scalability, they inevitably expand the attack surface and challenge cyber resilience. Reinforcement Learning (RL)-based defense strategies have been developed to optimize resource deployment and isolation policies under adversarial conditions, aiming to enhance system resilience by maintaining and restoring network availability. However, existing approaches lack robustness as they require retraining to adapt to dynamic changes in network structure, node scale, attack strategies, and attack intensity. Furthermore, the lack of Human-in-the-Loop (HITL) support limits interpretability and flexibility. To address these limitations, we propose CyberOps-Bots, a hierarchical multi-agent reinforcement learning framework empowered by Large Language Models (LLMs). Inspired by MITRE ATT&CK's Tactics-Techniques model, CyberOps-Bots features a two-layer architecture: (1) An upper-level LLM agent with four modules--ReAct planning, IPDRR-based perception, long-short term memory, and action/tool integration--performs global awareness, human intent recognition, and tactical planning; (2) Lower-level RL agents, developed via heterogeneous separated pre-training, execute atomic defense actions within localized network regions. This synergy preserves LLM adaptability and interpretability while ensuring reliable RL execution. Experiments on real cloud datasets show that, compared to state-of-the-art algorithms, CyberOps-Bots maintains network availability 68.5% higher and achieves a 34.7% jumpstart performance gain when shifting the scenarios without retraining. To our knowledge, this is the first study to establish a robust LLM-RL framework with HITL support for cloud defense. We will release our framework to the community, facilitating the advancement of robust and autonomous defense in cloud networks.

Yingjian Chen, Feiyang Li, Xingyu Song et al.

Large language models (LLMs) perform well in medical QA, but their effectiveness in Japanese contexts is limited due to privacy constraints that prevent the use of commercial models like GPT-4 in clinical settings. As a result, recent efforts focus on instruction-tuning open-source LLMs, though the potential of combining them with retrieval-augmented generation (RAG) remains underexplored. To bridge this gap, we are the first to explore a knowledge graph-based (KG) RAG framework for Japanese medical QA small-scale open-source LLMs. Experimental results show that KG-based RAG has only a limited impact on Japanese medical QA using small-scale open-source LLMs. Further case studies reveal that the effectiveness of the RAG is sensitive to the quality and relevance of the external retrieved content. These findings offer valuable insights into the challenges and potential of applying RAG in Japanese medical QA, while also serving as a reference for other low-resource languages.

Feiyang Li, Zinan Cao, Qixuan Yu et al.

For this paper, a prediction study of cloud computing energy consumption was conducted by optimising the data regression algorithm based on the horned lizard optimisation algorithm for Convolutional Neural Networks-Bi-Directional Gated Recurrent Units. Firstly, through Spearman correlation analysis of CPU, usage, memory usage, network traffic, power consumption, number of instructions executed, execution time and energy efficiency, we found that power consumption has the highest degree of positive correlation with energy efficiency, while CPU usage has the highest degree of negative correlation with energy efficiency. In our experiments, we introduced a random forest model and an optimisation model based on the horned lizard optimisation algorithm for testing, and the results show that the optimisation algorithm has better prediction results compared to the random forest model. Specifically, the mean square error (MSE) of the optimisation algorithm is 0.01 smaller than that of the random forest model, and the mean absolute error (MAE) is 0.01 smaller than that of the random forest.3 The results of the combined metrics show that the optimisation algorithm performs more accurately and reliably in predicting energy efficiency. This research result provides new ideas and methods to improve the energy efficiency of cloud computing systems. This research not only expands the scope of application in the field of cloud computing, but also provides a strong support for improving the energy use efficiency of the system.

Qixuan Yu, Xirui Tang, Feiyang Li et al.

This paper explores an improved Adaboost algorithm based on Long Short-Term Memory Networks (LSTMs), which aims to improve the prediction accuracy of user clicks on web page advertisements. By comparing it with several common machine learning algorithms, the paper analyses the advantages of the new model in ad click prediction. It is shown that the improved algorithm proposed in this paper performs well in user ad click prediction with an accuracy of 92%, which is an improvement of 13.6% compared to the highest of 78.4% among the other three base models. This significant improvement indicates that the algorithm is more capable of capturing user behavioural characteristics and time series patterns. In addition, this paper evaluates the model's performance on other performance metrics, including accuracy, recall, and F1 score. The results show that the improved Adaboost algorithm based on LSTM is significantly ahead of the traditional model in all these metrics, which further validates its effectiveness and superiority. Especially when facing complex and dynamically changing user behaviours, the model is able to better adapt and make accurate predictions. In order to ensure the practicality and reliability of the model, this study also focuses on the accuracy difference between the training set and the test set. After validation, the accuracy of the proposed model on these two datasets only differs by 1.7%, which is a small difference indicating that the model has good generalisation ability and can be effectively applied to real-world scenarios.

Yanming Mu, Hao Hu, Feiyang Li et al.

Retrieval-Augmented Generation (RAG) significantly mitigates the hallucinations and domain knowledge deficiency in large language models by incorporating external knowledge bases. However, the multi-module architecture of RAG introduces complex system-level security vulnerabilities. Guided by the RAG workflow, this paper analyzes the underlying vulnerability mechanisms and systematically categorizes core threat vectors such as data poisoning, adversarial attacks, and membership inference attacks. Based on this threat assessment, we construct a taxonomy of RAG defense technologies from a dual perspective encompassing both input and output stages. The input-side analysis reviews data protection mechanisms including dynamic access control, homomorphic encryption retrieval, and adversarial pre-filtering. The output-side examination summarizes advanced leakage prevention techniques such as federated learning isolation, differential privacy perturbation, and lightweight data sanitization. To establish a unified benchmark for future experimental design, we consolidate authoritative test datasets, security standards, and evaluation frameworks. To the best of our knowledge, this paper presents the first end-to-end survey dedicated to the security of RAG systems. Distinct from existing literature that isolates specific vulnerabilities, we systematically map the entire pipeline-providing a unified analysis of threat models, defense mechanisms, and evaluation benchmarks. By enabling deep insights into potential risks, this work seeks to foster the development of highly robust and trustworthy next-generation RAG systems.

Feiyang Li, Yingjian Chen, Haoran Liu et al.

Large Language Models (LLMs) have shown remarkable progress in medical question answering (QA), yet their effectiveness remains predominantly limited to English due to imbalanced multilingual training data and scarce medical resources for low-resource languages. To address this critical language gap in medical QA, we propose Multilingual Knowledge Graph-based Retrieval Ranking (MKG-Rank), a knowledge graph-enhanced framework that enables English-centric LLMs to perform multilingual medical QA. Through a word-level translation mechanism, our framework efficiently integrates comprehensive English-centric medical knowledge graphs into LLM reasoning at a low cost, mitigating cross-lingual semantic distortion and achieving precise medical QA across language barriers. To enhance efficiency, we introduce caching and multi-angle ranking strategies to optimize the retrieval process, significantly reducing response times and prioritizing relevant medical knowledge. Extensive evaluations on multilingual medical QA benchmarks across Chinese, Japanese, Korean, and Swahili demonstrate that MKG-Rank consistently outperforms zero-shot LLMs, achieving maximum 35.03% increase in accuracy, while maintaining an average retrieval time of only 0.0009 seconds.

Meilong Xu, Di Fu, Jiaxing Zhang et al.

Vision Language Models (VLMs) are becoming increasingly integral to multimedia understanding; however, they often struggle with domain-specific video classification tasks, particularly in cases with limited data. This stems from a critical \textit{rationale gap}, where sparse domain data is insufficient to bridge the semantic distance between complex spatio-temporal content and abstract classification labels. We propose a two-stage self-improvement paradigm to bridge this gap without new annotations. First, we prompt the VLMs to generate detailed textual rationales for each video, compelling them to articulate the domain-specific logic. The VLM is then fine-tuned on these self-generated rationales, utilizing this intermediate supervision to align its representations with the nuances of the target domain. Second, conventional supervised fine-tuning (SFT) is performed on the task labels, achieving markedly higher effectiveness as a result of the model's pre-acquired domain reasoning. Extensive experiments on diverse datasets demonstrate that our method significantly outperforms direct SFT, validating self-generated rationale as an effective, annotation-efficient paradigm for adapting VLMs to domain-specific video analysis.

Xirui Tang, Feiyang Li, Zinan Cao et al.

In this paper, an improved method for VR user experience prediction is investigated by introducing a sparrow search algorithm and a random forest algorithm improved by an iterative local search-optimised sparrow search algorithm. The study firstly conducted a statistical analysis of the data, and then trained and tested using the traditional random forest model, the random forest model improved by the sparrow search algorithm, and the random forest algorithm improved based on the iterative local search-sparrow search algorithm, respectively. The results show that the traditional random forest model has a prediction accuracy of 93% on the training set but only 73.3% on the test set, which is poor in generalisation; whereas the model improved by the sparrow search algorithm has a prediction accuracy of 94% on the test set, which is improved compared with the traditional model. What is more noteworthy is that the improved model based on the iterative local search-sparrow search algorithm achieves 100% accuracy on both the training and test sets, which is significantly better than the other two methods. These research results provide new ideas and methods for VR user experience prediction, especially the improved model based on the iterative local search-sparrow search algorithm performs well and is able to more accurately predict and classify the user's VR experience. In the future, the application of this method in other fields can be further explored, and its effectiveness can be verified through real cases to promote the development of AI technology in the field of user experience.