Yifan Hu

Relation Classification (RC) is a fundamental task in knowledge discovery. Prototypical networks are commonly used for few-shot RC tasks to tackle labeled data scarcity and long-tail relations, but they always overlook classification reliability and outlier features, potentially leading to sub-optimal similarity measurement. In a dynamic world with emerging novel semantic relations, incremental few-shot learning of relations gains attention, involving learning of both base and novel relations in new tasks while retaining base relation knowledge. However, this becomes challenging with increasing novel relations, requiring that models should reduce reliance on learning base relations for the new task. Therefore, this article explores Class-Incremental Few-Shot Relation Classification (CIFRC) using a Gaussian-Augmented Prototypical Network (GA-Proto). GA-Proto refines the similarity measurement by incorporating discrepancies between ideal and actual classifications via Gaussian Mixture Model and analyzing Gaussian outlier features of query instances. It also employs reliability learning and knowledge distillation to mitigate encoding space distortion and base relation forgetting by enhancing classification reliability and transferring base relation knowledge, respectively. Additionally, GA-Proto uses label smoothing to alleviate novel relation overfitting. Experimental results on three public datasets demonstrate that GA-Proto outperforms existing methods on CIFRC, achieving up to 19.64% improvement in accuracy. The datasets and source code for GA-Proto are released at https://github.com/GTZN2/GA-Proto .

Large language models (LLMs) exhibit exceptional linguistic capabilities, yet their ability to verbatim reproduce copyrighted data from high-quality training datasets raises concerns about improper exploitation via Artificial Intelligence (AI) data traceability. Such copyrighted data may include general information (e.g., historical knowledge, or common sense), necessitating that LLMs generate content preserving key information rather than exact replicas. Current unlearning methods, however, primarily aim to fully forget targeted knowledge or key information, potentially leading to hallucinations of general information. To address these challenges, we propose an unlearning method based on attention flattening in auto-regressive models, combined with simulation tasks for targeted information forgetting. During optimization, the model is trained only on real tasks while acquiring knowledge from simulation task learning to suppress the memorization of copyrighted data. In inference, we compare the sensitivity of attention heads between simulation and real tasks, identify those most relevant to copyrighted data, and perform attention head pruning to suppress the generation of such data. Experimental results show that our method achieves lower perplexity and BLEU scores while maintaining higher entity coverage scores compared to existing unlearning baselines in generation tasks on copyrighted general information.

Static characteristics of supervisory control and data acquisition (SCADA) system are often exploited to perform malicious activities on smart grids. Most of the time, the success of cyberattacks begins with the profiling of the target system and follows by the analysis of the limited resources. To alleviate the asymmetry between attack and defense, network-based moving target defense (MTD) techniques have been applied in the network system to defend against cyberattacks by constructing a dynamic attack surface to the adversary. In this paper, we propose a novel MTD technique based on adaptive forwarding path migration (AFPM) that focuses on improving the defense capability and optimizing the network performance of path mutation. Considering the transient problems present in path mutation caused by the dynamic switching of the forwarding path, we formalize the mutation constraints based on the satisfiability modulo theory (SMT) to select the mutation path. Considering the limited defense capability of path mutation owing to the traditional mutation selection mechanism, we design the mutation path generation algorithm based on the network security capacity matrix to obtain an optimal combination of mutation path and mutation period. Finally, we compare and analyze various cyber defense techniques used in the SCADA network and demonstrate experimentally that our MTD technique can prevent more than 92% of passive monitoring under specified conditions while ensuring the quality of service (QoS) to be almost the same as the static network.

The ever-expanding power system is developed into an interconnected pattern of power grids. Zone partitioning is an essential technique for the operation and management of such an interconnected power system. Owing to the transmission capacity limitation, transmission congestion may occur with a regional influence on power system. If transmission congestion is considered when the system is decomposed into several regions, the power consumption structure can be optimized and power system planning can be more reasonable. At the same time, power resources can be properly allocated and system safety can be improved. In this paper, we propose a power system zone partitioning method where the potential congested branches are identified and the spectral clustering algorithm is improved. We transform the zone partitioning problem into a graph segmentation problem by constructing an undirected weighted graph of power system where the similarities between buses are measured by the power transfer distribution factor (PTDF) corresponding to the potential congested branches. Zone partitioning results show that the locational marginal price (LMP) in the same zone is similar, which can represent regional price signals and provide regional auxiliary decisions.

This paper presents a novel parallel quasi-cyclic low-density parity-check (QC-LDPC) encoding algorithm with low complexity, which is compatible with the 5th generation (5G) new radio (NR). Basing on the algorithm, we propose a high area-efficient parallel encoder with compatible architecture. The proposed encoder has the advantages of parallel encoding and pipelined operations. Furthermore, it is designed as a configurable encoding structure, which is fully compatible with different base graphs of 5G LDPC. Thus, the encoder architecture has flexible adaptability for various 5G LDPC codes. The proposed encoder was synthesized in a 65 nm CMOS technology. According to the encoder architecture, we implemented nine encoders for distributed lifting sizes of two base graphs. The eperimental results show that the encoder has high performance and significant area-efficiency, which is better than related prior art. This work includes a whole set of encoding algorithm and the compatible encoders, which are fully compatible with different base graphs of 5G LDPC codes. Therefore, it has more flexible adaptability for various 5G application scenarios.

A large-scale Cyber-Physical System (CPS) such as a smart grid usually provides service to a vast number of users as a public utility. Security is one of the most vital aspects in such critical infrastructures. The existing CPS security usually considers the attack from the information domain to the physical domain, such as injecting false data to damage sensing. Social Collective Attack on CPS (SCAC) is proposed as a new kind of attack that intrudes into the social domain and manipulates the collective behavior of social users to disrupt the physical subsystem. To provide a systematic description framework for such threats, we extend MITRE ATT&CK, the most used cyber adversary behavior modeling framework, to cover social, cyber, and physical domains. We discuss how the disinformation may be constructed and eventually leads to physical system malfunction through the social-cyber-physical interfaces, and we analyze how the adversaries launch disinformation attacks to better manipulate collective behavior. Finally, simulation analysis of SCAC in a smart grid is provided to demonstrate the possibility of such an attack.