XLXinghao Li

Xinghao Li

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Xinghao Li has a total of 16 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Logits-Level Balanced Machine Unlearning for LLM-Based Recommendation System

    by Chenchen Tan, Xinghao Li, Youyang Qu, Cunjian Chen, Shujie Cui, Longxiang Gao

    As recommendation systems have become foundational to digital platforms, the integration of large language models (LLMs) into these systems offers transformative potential. In LLM-based recommendation (LLMRec) systems, LLMs excel at capturing nuanced user pReferences, greatly enhancing personalization and recommendation accuracy. However, LLMRec systems face data governance issues, such as data privacy, outdated data, poisoned data, and copyrighted data. All these data government scenarios require the removal of data and its impact from large models like LLMRec. This poses significant challenges as existing solutions struggle to accurately delete target data from such complex systems. Motivated by this, we propose an LLMRec unlearning system based on adapter-driven logits modification in this work. This mechanism addresses the unique characteristics of LLMRec’s complex parameters and network structure, enabling precise and effective unlearning while maintaining recommendation performance. Our approach offers several key advantages: 1) using adapters reduces training costs during the unlearning process; 2) a knowledge distillation (KD)-powered logits modification mechanism ensures that the model can still engage in effective reasoning after unlearning, leveraging its existing knowledge logic to maintain high-quality recommendations; and 3) an additional adapter-based general knowledge retention module is incorporated to preserve the core language capabilities of the LLMRec system, ensuring that its foundational performance remains intact. Extensive experimental results show the effectiveness and efficiency of the proposed system. The code is released at: https://anonymous.4open.science/r/llmrecunlearning-1CB3


  • Navigating Unlearning in Medical AI: A Framework for Diabetic Retinopathy Classification

    by Xinghao Li, Chi Liu, Youyang Qu, Shujie Cui, Cunjian Chen, Xingliang Yuan, Zongyuan Ge, Longxiang Gao

    Medical Artificial Intelligence (AI) systems hold great potential for clinical and diagnostic applications. However, their deployment raises significant privacy and ethical concerns, primarily due to the reliance on large-scale datasets containing sensitive health information. The processes of collecting, storing, and using patient data bring forth widespread concerns about privacy protection and data ownership. Additionally, the dynamic nature of medical data presents risks related to outdated or inaccurate information, which can directly impact patient care. Introducing the concept of “unlearning” in medical AI is therefore crucial, not only to address patient privacy needs but also to ensure that models remain adaptable to evolving medical knowledge and data quality. Nevertheless, implementing unlearning in medical AI is considerably more complex compared to other domains, as errors in this context could lead to severe consequences for patient health. This paper aims to establish an evaluation framework specifically for unlearning in classification tasks on the Mobile Brazilian Retinal Dataset (mBRSET) in medical AI. The framework addresses key issues such as privacy, clinical safety, diagnostic accuracy, and fairness. By providing a detailed foundation for assessing unlearning practices, the framework is tailored to the unique challenges presented by the mBRSET dataset. Ultimately, this work seeks to identify the trade-offs and requirements necessary to implement responsible unlearning processes in this highly sensitive domain, thereby contributing to the safe and ethical advancement of medical AI.

  • Breaking Down Data Sharing Barrier of Smart City: A Digital Twin Approach

    by Guanjie Li, Tom H. Luan, Xinghao Li, Jinkai Zheng, Chengzhe Lai, Zhou Su, Kuan Zhang

    The smart city aims to enhance the efficacy of urban management and improve the quality of citizen life. This however faces a fundamental difficulty due to the inadequate data sharing among social entities, which has become one of the key barriers to the in-depth development of smart city, making wasted information resources and application silos ubiquitous. As an emerging communication paradigm, the digital twin network (DTN) has the potential to eliminate the data-sharing barrier in smart cities. By mapping the social physical entities (PEs) in the smart city to digital virtual agents, namely data twins (DTs), valuable data can be directly shared among the DTs over DTN without physical constraints, and form decisions that feedback and guide the operations of PEs. In this article, we explore the role of DTN in supporting the smart city and promoting data-sharing construction. We first introduce the basic concepts related to DTN and propose the architecture of DTN-enabled smart city. The key features of the DTN-enabled smart city are then discussed with an emphasis on the security and privacy issues with fundamental solutions around data sharing in the DTN-enabled smart city. As a case study, we investigate data sharing in the DTN-enabled smart city and design a secure and fair data-sharing mechanism. Finally, we propose open challenges for achieving a more reliable DTN-enabled smart city.


  • Digital Twin Enabled Remote Data Sharing for Internet of Vehicles: System and Incentive Design

    by Chenchen Tan, Xinghao Li, Longxiang Gao, Tom H. Luan, Youyang Qu, Yong Xiang, Rongxing Lu

    With the boom in advanced intelligent vehicles, the amount of data associated with the Internet of Vehicles (IoVs) grows exponentially. Sharing road data among vehicles effectively can greatly improve their driving efficiency and road experience. This, however, is hindered by security concerns, bandwidth limitations, and mobility and communication distance between vehicles. In this article, we propose a citywide data sharing platform among vehicles based on the digital twin. Specifically, we propose a digital twin based vehicular platform on the cloud to facilitate the effective data sharing of vehicles. A digital twin is a digital representation of a vehicle on the cloud that synchronises data with the vehicle in real time. Therefore, data sharing among vehicles can be accomplished by their digital twins on the cloud without any physical limitations. Considering the data privacy of vehicles, a framework that combines federated learning and transfer learning is applied which can realise personalised data sharing among vehicles without disclosing their data privacy. An incentive mechanism based on the game-theoretic approach is devised to combat the mutual distrust between vehicles and encourage their contribution to data sharing. By sewing the above mechanisms, the proposed digital twin enabled data sharing platform can effectively address privacy, bandwidth limitation and incentive issues. Using extensive trace-driven simulations, we demonstrate the effectiveness and efficiency of the proposed system.