Author name cluster

Weihua Li

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

11 papers

2 author rows

AAAI Conference 2026 Conference Paper

FedSEA-LLaMA: A Secure, Efficient and Adaptive Federated Splitting Framework for Large Language Models

Zishuai Zhang
Hainan Zhang
Weihua Li
Qinnan Zhang
Jin Dong
Yongxin Tong
Zhiming Zheng

Private data holds promise for improving LLMs due to its high quality, but its scattered distribution across data silos and the high computational demands of LLMs limit their deployment in federated environments. To address this, the transformer-based federated split models are proposed, which offload most model parameters to the server (or distributed clients) while retaining only a small portion on the client to ensure data privacy. Despite this design, they still face three challenges: 1) Peer-to-peer key encryption struggles to secure transmitted vectors effectively; 2) The auto-regressive nature of LLMs means that federated split learning can only train and infer sequentially, causing high communication overhead; 3) Fixed partition points lack adaptability to downstream tasks. In this paper, we introduce FedSEA-LLaMA, a Secure, Efficient, and Adaptive Federated splitting framework based on LLaMA2. First, we inject Gaussian noise into forward-pass hidden states to enable secure end-to-end vector transmission. Second, we employ attention-mask compression and KV cache collaboration to reduce communication costs, accelerating training and inference. Third, we allow users to dynamically adjust the partition points for input/output blocks based on specific task requirements. Experiments on natural language understanding, summarization, and conversational QA tasks show that FedSEA-LLaMA maintains performance comparable to centralized LLaMA2 and achieves up to 8× speedups in training and inference. Further analysis of privacy attacks and different partition points also demonstrates the effectiveness of FedSEA-LLaMA in security and adaptability.

PDF Details DOI

AIIM Journal 2025 Journal Article

Enhancing diagnosis prediction with adaptive disease representation learning

Hengliang Cheng
Shibo Li
Tao Shen
Weihua Li

Diagnosis prediction predicts which diseases a patient is most likely to suffer from in the future based on their historical electronic health records. The time series model can better capture the temporal progression relationship of patient diseases, but ignores the semantic correlation between all diseases; in fact, multiple diseases that are often diagnosed at the same time reflect hidden patterns that are conducive to diagnosis, so predefined global disease co-occurrence graph can help the model understand disease relationships. But it may contain a lot of noise and ignore the semantic adaptation of the disease under the diagnosis target. To this end, we propose a graph-driven end-to-end framework, named A daptive D isease R epresentation L earning (ADRL), obtain disease representation after learning complex disease relationships, and then use it to improve diagnosis prediction performance. This model introduces an adaptive mechanism to dynamically adjust and optimize disease relationships by performing self-supervised perturbations on a predefined global disease co-occurrence graph, thereby learning a global disease relationship graph that contains complex semantic association information between diseases. The computational burden of adaptive global disease graph can be further alleviated by the proposed SVD-based accelerator. Finally, experimental results on two real-world EHR datasets show that the proposed model outperforms existing models in diagnosis prediction.

Details DOI

EAAI Journal 2024 Journal Article

A digital twin-driven approach for partial domain fault diagnosis of rotating machinery

Jingyan Xia
Zhuyun Chen
Jiaxian Chen
Guolin He
Ruyi Huang
Weihua Li

Artificial intelligence (AI)-driven fault diagnosis methods are crucial for ensuring rotating machinery's safety and effective operation. The success of most current methods relies on the assumption that sufficient high-quality labeled datasets can be obtained for model training. However, in real-world industrial scenarios, obtaining such datasets is difficult or nearly impossible, thereby hindering the practical implementation of these methods. The integration of virtual modeling and transfer learning offers a powerful approach to meet the above challenge. Abundant virtual data of different fault categories can be acquired in the virtual space with highly flexible and at a low cost, and transfer learning can enhance the practical utility of these virtual data for contributing to the construction of diagnosis models. Therefore, this paper proposes a digital twin-driven partial domain fault diagnosis method based on unlabeled physical data and labeled virtual data. First, a virtual model of rotating machinery is built to generate labeled virtual fault data with enough fault types. Then, an adversarial transfer learning network is developed to leverage the effective knowledge from the virtual and physical data. Meanwhile, a weighting learning module is introduced to reduce the negative effect caused by the redundant fault categories in the virtual space. Finally, the proposed digital twin-driven transfer learning network is trained with the labeled virtual data and unlabeled physical data. Experiments on a light truck transmission system demonstrate that the proposed method achieves satisfactory diagnostic performance even without labeled physical fault data, contributing to the advancement of AI engineering applications.

Details DOI

IJCAI Conference 2024 Conference Paper

An LLM-enhanced Agent-based Simulation Tool for Information Propagation

Yuxuan Hu
Gemju Sherpa
Lan Zhang
Weihua Li
Quan Bai
Yijun Wang
Xiaodan Wang

Influence diffusion models are used for simulating information propagation in social networks. While most existing influence diffusion models are probabilistic, the emergence of Large Language Model (LLM) sheds light on the language-level inferences and interactions of user agents. This paper presents an LLM-enhanced Agent-based Influence Diffusion model (LAID), and a web-based visualization tool, LAIDSim, for simulating the information propagation in social networks.

PDF Details DOI

ICLR Conference 2023 Conference Paper

Simple and Scalable Nearest Neighbor Machine Translation

Yuhan Dai
Zhirui Zhang
Qiuzhi Liu
Qu Cui
Weihua Li
Yichao Du
Tong Xu 0001

$k$NN-MT is a straightforward yet powerful approach for fast domain adaptation, which directly plugs the pre-trained neural machine translation (NMT) models with domain-specific token-level $k$-nearest-neighbor ($k$NN) retrieval to achieve domain adaptation without retraining. Despite being conceptually attractive, $k$NN-MT is burdened with massive storage requirements and high computational complexity since it conducts nearest neighbor searches over the entire reference corpus. In this paper, we propose a simple and scalable nearest neighbor machine translation framework to drastically promote the decoding and storage efficiency of $k$NN-based models while maintaining the translation performance. To this end, we dynamically construct a extremely small datastore for each input via sentence-level retrieval to avoid searching the entire datastore in vanilla $k$NN-MT, based on which we further introduce a distance-aware adapter to adaptively incorporate the $k$NN retrieval results into the pre-trained NMT models. Experiments on machine translation in two general settings, static domain adaptation, and online learning, demonstrate that our proposed approach not only achieves almost 90% speed as the NMT model without performance degradation, but also significantly reduces the storage requirements of $k$NN-MT.

Details

IJCAI Conference 2022 Conference Paper

AI Facilitated Isolations? The Impact of Recommendation-based Influence Diffusion in Human Society

Yuxuan Hu
Shiqing Wu
Chenting Jiang
Weihua Li
Quan Bai
Erin Roehrer

AI recommendation techniques provide users with personalized services, feeding them the information they may be interested in. The increasing personalization raises the hypotheses of the "filter bubble" and "echo chamber" effects. To investigate these hypotheses, in this paper, we inspect the impact of recommendation algorithms on forming two types of ideological isolation, i. e. , the individual isolation and the topological isolation, in terms of the filter bubble and echo chamber effects, respectively. Simulation results show that AI recommendation strategies severely facilitate the evolution of the filter bubble effect, leading users to become ideologically isolated at an individual level. Whereas, at a topological level, recommendation algorithms show eligibility in connecting individuals with dissimilar users or recommending diverse topics to receive more diverse viewpoints. This research sheds light on the ability of AI recommendation strategies to temper ideological isolation at a topological level.

PDF Details DOI

AAMAS Conference 2021 Conference Paper

Graph-based Self-Adaptive Conversational Agent

Lan Zhang
Weihua Li
Quan Bai
Edmund Lai

Conversational agents have been widely adopted in dialogue systems for various business purposes. Many existing conversational agents are rule-based and require significant human intervention to adapt the knowledge and conversational flow. In this paper, we propose a graph-based adaptive conversational agent model which is capable of learning knowledge from human beings and adapting the knowledge-base according to human-agent interactions. Studies to evaluate the proposed model are conducted and presented, which compare the responses from the proposed adaptive agent model and a conventional agent.

PDF

AAMAS Conference 2021 Conference Paper

Learning Policies for Effective Incentive Allocation in Unknown Social Networks

Shiqing Wu
Quan Bai
Weihua Li

Most existing incentive allocation approaches rely on sufficient information about users’ attributes, such as their preferences, followers in the social network, and activities, to customize effective incentives. However, this may lead to failure when such knowledge is unavailable. In this light, we propose an end-to-end reinforcement learning-based framework, named Geometric Actor-Critic (GAC), to discover effective incentive allocation policies towards users in a social network. More specifically, given a limited budget, the proposed approach can extract information from a high-level network representation for learning effective incentive allocation policies. The proposed GAC only requires the topology of the social network and does not rely on any prior information about users’ attributes. We use three real-world social network datasets to evaluate the performance of the proposed GAC. The experimental results demonstrate the effectiveness of the proposed approach.

PDF

YNICL Journal 2020 Journal Article

Longitudinal functional connectivity changes related to dopaminergic decline in Parkinson’s disease

Weihua Li
Nick P. Lao-Kaim
Andreas-Antonios Roussakis
Antonio Martín-Bastida
Natalie Valle-Guzman
Gesine Paul
Eyal Soreq
Richard E. Daws

BACKGROUND: Resting-state functional magnetic resonance imaging (fMRI) studies have demonstrated that basal ganglia functional connectivity is altered in Parkinson's disease (PD) as compared to healthy controls. However, such functional connectivity alterations have not been related to the dopaminergic deficits that occurs in PD over time. OBJECTIVES: To examine whether functional connectivity impairments are correlated with dopaminergic deficits across basal ganglia subdivisions in patients with PD both cross-sectionally and longitudinally. METHODS: ) was calculated for each participant. PD patients were assessed for disease severity. RESULTS: C-PE2I PET, showed larger decreases in posterior putamen functional connectivity with the midbrain and pallidum. Reduced functional connectivity of the posterior putamen with the thalamus, midbrain, supplementary motor area and sensorimotor cortex over time were significantly associated with changes in DAT density over the same period. Furthermore, increased motor disability was associated with lower intraregional functional connectivity of the posterior putamen. CONCLUSIONS: Our findings suggest that basal ganglia functional connectivity is related to integrity of dopaminergic system in patients with PD. Application of resting-state fMRI in a large cohort and longitudinal scanning may be a powerful tool for assessing underlying PD pathology and its progression.

Details DOI

AAMAS Conference 2018 Conference Paper

Modelling Multiple Influences Diffusion in On-line Social Networks

Weihua Li
Quan Bai
Minjie Zhang
Tung Doan Nguyen

In on-line social networks, innovations in the presence of one or more influences disseminate through the topological structure of the networks rapidly. In reality, various influences normally coexist in the same context and have subtle relations, such as supportive, contradictory and competitive relations, affecting the users’ decisions of adopting any innovations. Therefore, modelling diffusion process of multiple influences is an important, yet challenging research question. By employing the agent-based modelling, in this paper, a distributed approach has been proposed to model the diffusion process of multiple influences in social networks. The proposed model has been applied in the undesirable influence minimisation problem, where the time series is taken into consideration. The experimental results show our model can be utilised to minimise the adverse impact of a certain influence by injecting other influences. Furthermore, the proposed model also sheds light on understanding, investigating and analysing multiple influences in social networks

PDF

AAMAS Conference 2017 Conference Paper

Agent-based Influence Maintenance in Social Networks

Weihua Li
Quan Bai
Tung Doan Nguyen
Minjie Zhang

We study on how to maintain long-term influence in a social network by proposing an agent-based influence maintenance model. Within the context of our investigation, the experimental results reveal that multiple-time seed selection is capable of achieving more constant impact than one-shot selection.

PDF