Yisheng Zhong

Zhuang Jiao, Yisheng Zhong

This paper has been withdrawn. This paper focuses on the admissibility condition for fractional-order singular system with order $α\in (0,1)$. The definitions of regularity, impulse-free and admissibility are given first, then a sufficient and necessary condition of admissibility for fractional-order singular system is established. A numerical example is included to illustrate the proposed condition.

Zhuang Jiao, Yisheng Zhong

The issues of robust stability for two types of uncertain fractional-order systems of order $α\in (0,1)$ are dealt with in this paper. For the polytope-type uncertainty case, a less conservative sufficient condition of robust stability is given; for the norm-bounded uncertainty case, a sufficient and necessary condition of robust stability is presented. Both of these conditions can be checked by solving sets of linear matrix inequalities. Two numerical examples are presented to confirm the proposed conditions.

Rui Yan, Zongying Shi, Yisheng Zhong

This paper considers a subspace guarding game in high-dimensional space which consists of a play subspace and a target subspace. Two faster defenders cooperate to protect the target subspace by capturing an attacker which strives to enter the target subspace from the play subspace without being captured. A closed-form solution is provided from the perspectives of kind and degree. Contributions of the work include the use of the attack subspace (AS) method to construct the barrier, by which the game winner can be perfectly predicted before the game starts. In addition to this inclusion, with the priori information about the game result, a critical payoff function is designed when the defenders can win the game. Then, the optimal strategy for each player is explicitly reformulated as a saddle-point equilibrium. Finally, we apply these theoretical results to a half-space guarding game in three-dimensional space. Since the whole achieved developments are analytical, they require a little memory without computational burden and allow for real-time updates, beyond the capacity of traditional Hamilton-Jacobi-Isaacs method. It is worth noting that this is the first time in the current work to consider the target guarding games for arbitrary high-dimensional space, and in a fully analytical form.

Yisheng Zhong, Zhengbang Yang, Zhuangdi Zhu

LLM unlearning is a technique to remove the impacts of undesirable knowledge from the model without retraining from scratch, which is indispensable towards trustworthy AI. Existing unlearning methods face significant limitations: conventional tuning-based unlearning is computationally heavy and prone to catastrophic forgetting. In contrast, in-contextualized unlearning is lightweight for precise unlearning but vulnerable to prompt removal or reverse engineering attacks. In response, we propose Distilled Unlearning from an Efficient Teacher (DUET), a novel distillation-based unlearning method that combines the merits of these two lines of work. It learns a student model to imitate the behavior of a prompt-steered teacher that effectively refuses undesirable knowledge generation while preserving general domain knowledge. Extensive evaluations on existing benchmarks with our enriched evaluation protocols demonstrate that DUET achieves higher performance in both forgetting and utility preservation, while being orders of magnitude more data-efficient than state-of-the-art unlearning methods.

Yisheng Zhong, Sijia Liu, Zhuangdi Zhu

Large Language Models (LLMs) unlearning is crucial for removing hazardous or privacy-leaking information from the model. Practical LLM unlearning demands satisfying multiple challenging objectives simultaneously: removing undesirable knowledge, preserving general utility, avoiding over-refusal of neighboring concepts, and, crucially, ensuring robustness against adversarial probing attacks. However, existing unlearning methods primarily focus on a limited subset of these goals, typically unlearning efficacy and utility preservation while overlooking robustness and boundary behaviors. Naively extending these methods to multi-objective settings may lead to unlearning task interference. We propose a novel multi-objective unlearning framework that harmonizes multiple unlearning objectives through a data and optimization co-design: We standardize training corpora into a unified data representation to reduce the domain gap, and then introduce a bidirectional distillation method that simultaneously elicits desired behavior from a context-instructed teacher while suppressing undesirable behavior in the student model. Theoretical and empirical analyses show that our method aligns domain distributions and converts seemingly irrelevant unlearning tasks into cooperative optimization. Evaluation demonstrates state-of-the-art performance, which enables balanced and reliable unlearning across diverse, challenging requirements.

Zhengbang Yang, Yisheng Zhong, Junyuan Hong et al.

Pretrained knowledge memorized in LLMs raises critical concerns over safety and privacy, which has motivated LLM Unlearning as a technique for selectively removing the influences of undesirable knowledge. Existing approaches, rooted in Gradient Ascent (GA), often degrade general domain knowledge while relying on retention data or curated contrastive pairs, which can be either impractical or data and computationally prohibitive. Negative Preference Alignment has been explored for unlearning to tackle the limitations of GA, which, however, remains confined by its choice of reference model and shows undermined performance in realistic data settings. These limitations raise two key questions: i) Can we achieve effective unlearning that quantifies model confidence in undesirable knowledge and uses it to calibrate gradient updates more precisely, thus reducing catastrophic forgetting? ii) Can we make unlearning robust to data scarcity and length variation? We answer both questions affirmatively with CATNIP (Calibrated and Tokenized Negative Preference Alignment), a principled method that rescales unlearning effects in proportion to the model's token-level confidence, thus ensuring fine-grained control over forgetting. Extensive evaluations on MUSE and WMDP benchmarks demonstrated that our work enables effective unlearning without requiring retention data or contrastive unlearning response pairs, with stronger knowledge forgetting and preservation tradeoffs than state-of-the-art methods.

Yisheng Zhong, Zhengbang Yang, Zhuangdi Zhu

Large Language Models (LLMs) are increasingly integrated into real-world applications, raising concerns about privacy, security and the need to remove undesirable knowledge. Machine Unlearning has emerged as a promising solution, yet faces two key challenges: (1) practical unlearning needs are often continuous and heterogeneous, and (2) they involve decentralized, sensitive data with asymmetric access. These factors result in inter-domain and intra-domain interference, which further amplifies the dilemma of unbalanced forgetting and retaining performance. In response, we propose a federated unlearning approach for LLMs that is scalable and privacy preserving. Our method decouples unlearning and retention via task-specific adapter learning and employs a hierarchical merging strategy to mitigate conflicting objectives and enables robust, adaptable unlearning updates. Comprehensive experiments on benchmarks of WMDP, MUSE, and TOFU showed that our approach effectively handles heterogeneous unlearning requests while maintaining strong LLM utility compared with baseline methods.

Yisheng Zhong, Yizhu Wen, Junfeng Guo et al.

The protection of cyber Intellectual Property (IP) such as web content is an increasingly critical concern. The rise of large language models (LLMs) with online retrieval capabilities enables convenient access to information but often undermines the rights of original content creators. As users increasingly rely on LLM-generated responses, they gradually diminish direct engagement with original information sources, which will significantly reduce the incentives for IP creators to contribute, and lead to a saturating cyberspace with more AI-generated content. In response, we propose a novel defense framework that empowers web content creators to safeguard their web-based IP from unauthorized LLM real-time extraction and redistribution by leveraging the semantic understanding capability of LLMs themselves. Our method follows principled motivations and effectively addresses an intractable black-box optimization problem. Real-world experiments demonstrated that our methods improve defense success rates from 2.5% to 88.6% on different LLMs, outperforming traditional defenses such as configuration-based restrictions.

Rui Yan, Xiaoming Duan, Zongying Shi et al.

This paper introduces two metrics (cycle-based and memory-based metrics), grounded on a dynamical game-theoretic solution concept called sink equilibrium, for the evaluation, ranking, and computation of policies in multi-agent learning. We adopt strict best response dynamics (SBRD) to model selfish behaviors at a meta-level for multi-agent reinforcement learning. Our approach can deal with dynamical cyclical behaviors (unlike approaches based on Nash equilibria and Elo ratings), and is more compatible with single-agent reinforcement learning than alpha-rank which relies on weakly better responses. We first consider settings where the difference between largest and second largest underlying metric has a known lower bound. With this knowledge we propose a class of perturbed SBRD with the following property: only policies with maximum metric are observed with nonzero probability for a broad class of stochastic games with finite memory. We then consider settings where the lower bound for the difference is unknown. For this setting, we propose a class of perturbed SBRD such that the metrics of the policies observed with nonzero probability differ from the optimal by any given tolerance. The proposed perturbed SBRD addresses the opponent-induced non-stationarity by fixing the strategies of others for the learning agent, and uses empirical game-theoretic analysis to estimate payoffs for each strategy profile obtained due to the perturbation.

Xiaolong Ma, Xiatian Zhu, Shaogang Gong et al.

Most existing person re-identification (ReID) methods rely only on the spatial appearance information from either one or multiple person images, whilst ignore the space-time cues readily available in video or image-sequence data. Moreover, they often assume the availability of exhaustively labelled cross-view pairwise data for every camera pair, making them non-scalable to ReID applications in real-world large scale camera networks. In this work, we introduce a novel video based person ReID method capable of accurately matching people across views from arbitrary unaligned image-sequences without any labelled pairwise data. Specifically, we introduce a new space-time person representation by encoding multiple granularities of spatio-temporal dynamics in form of time series. Moreover, a Time Shift Dynamic Time Warping (TS-DTW) model is derived for performing automatically alignment whilst achieving data selection and matching between inherently inaccurate and incomplete sequences in a unified way. We further extend the TS-DTW model for accommodating multiple feature-sequences of an image-sequence in order to fuse information from different descriptions. Crucially, this model does not require pairwise labelled training data (i.e. unsupervised) therefore readily scalable to large scale camera networks of arbitrary camera pairs without the need for exhaustive data annotation for every camera pair. We show the effectiveness and advantages of the proposed method by extensive comparisons with related state-of-the-art approaches using two benchmarking ReID datasets, PRID2011 and iLIDS-VID.