检索结果-内蒙古大学图书馆

A2PO: towards effective offline reinforcement learning from an advantage-aware perspective 24

学校读者我要写书评

暂无评论

A2PO: towards effective offline reinforcement learning from ...

Proceedings of the 38th International Conference on Neural Information Processing Systems

作者： Yunpeng Qing Shunyu Liu Jingyuan Cong Kaixuan Chen Yihe Zhou Mingli Song College of Computer Science and Technology Zhejiang University State Key Laboratory of Blockchain and Data Security Zhejiang University State Key Laboratory of Blockchain and Data Security Zhejiang University and Hangzhou High-Tech Zone (Binjiang) Institute of Blockchain and Data Security

ISBN: (纸本)9798331314385

Offline reinforcement learning endeavors to leverage offline datasets to craft effective agent policy without online interaction, which imposes proper conservative constraints with the support of behavior policies to tackle the out-of-distribution problem. However, existing works often suffer from the constraint conflict issue when offline datasets are collected from multiple behavior policies, i.e., different behavior policies may exhibit inconsistent actions with distinct returns across the state space. To remedy this issue, recent advantage-weighted methods prioritize samples with high advantage values for agent training while inevitably ignoring the diversity of behavior policy. In this paper, we introduce a novel Advantage-Aware Policy Optimization (A2PO) method to explicitly construct advantage-aware policy constraints for offline learning under mixed-quality datasets. Specifically, A2PO employs a conditional variational auto-encoder to disentangle the action distributions of intertwined behavior policies by modeling the advantage values of all training data as conditional variables. Then the agent can follow such disentangled action distribution constraints to optimize the advantage-aware policy towards high advantage values. Extensive experiments conducted on both the single-quality and mixed-quality datasets of the D4RL benchmark demonstrate that A2PO yields results superior to the counterparts. Our code is available at https://***/Plankson/A2PO.

关键词：

Can Small Language Models Reliably Resist Jailbreak Attacks? A Comprehensive Evaluation

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Zhang, Wenhui Xu, Huiyu Wang, Zhibo He, Zeqing Zhu, Ziqi Ren, Kui The State Key Laboratory of Blockchain and Data Security Zhejiang University China

Small language models (SLMs) have emerged as promising alternatives to large language models (LLMs) due to their low computational demands, enhanced privacy guarantees and comparable performance in specific domains through light-weight fine-tuning. Deploying SLMs on edge devices, such as smartphones and smart vehicles, has become a growing trend. However, the security implications of SLMs have received less attention than LLMs, particularly regarding jailbreak attacks, which is recognized as one of the top threats of LLMs by the OWASP. In this paper, we conduct the first large-scale empirical study of SLMs’ vulnerabilities to jailbreak attacks. Through systematically evaluation on 63 SLMs from 15 mainstream SLM families against 8 state-of-the-art jailbreak methods, we demonstrate that 47.6% of evaluated SLMs show high susceptibility to jailbreak attacks (ASR > 40%) and 38.1% of them can not even resist direct harmful query (ASR > 50%). We further analyze the reasons behind the vulnerabilities and identify four key factors: model size, model architecture, training datasets and training techniques. Moreover, we assess the effectiveness of three prompt-level defense methods and find that none of them achieve perfect performance, with detection accuracy varying across different SLMs and attack methods. Notably, we point out that the inherent security awareness play a critical role in SLM security, and models with strong security awareness could timely terminate unsafe response with little reminder. Building upon the findings, we highlight the urgent need for security-by-design approaches in SLM development and provide valuable insights for building more trustworthy SLM ecosystem. Copyright © 2025, The Authors. All rights reserved.

关键词： Structured Query Language

MUT: Human-in-the-Loop Unit Test Migration 24

学校读者我要写书评

暂无评论

MUT: Human-in-the-Loop Unit Test Migration

44th ACM/IEEE International Conference on Software Engineering, ICSE 2024

作者： Gao, Yi Hu, Xing Xu, Tongtong Xia, Xin Lo, David Yang, Xiaohu Zhejiang University The State Key Laboratory of Blockchain and Data Security Hangzhou China Huawei Hangzhou China School of Computing and Information Systems Singapore Management University Hangzhou China

ISBN: (纸本)9798400702174

Test migration, which enables the reuse of test cases crafted with knowledge and creativity by testers across various platforms and programming languages, has exhibited effectiveness in mobile app testing. However, unit test migration at the source code level has not garnered adequate attention and exploration. In this paper, we propose a novel cross-language and cross-platform test migration methodology, named MUT, which consists of four modules: code mapping, test case filtering, test case translation, and test case adaptation. MUT initially calculates code mappings to establish associations between source and target projects, and identifies suitable unit tests for migration from the source project. Then, MUT's code translation component generates a syntax tree by parsing the code to be migrated and progressively converts each node in the tree, ultima tely generating the target tests, which are compiled and executed in the target project. Moreover, we develop a web tool to assist developers in test migration. The effectiveness of our approach has been validated on five prevalent functional domain projects within the open-source community. We migrate a total of 550 unit tests and submitted pull requests to augment test code in the target projects on GitHub. By the time of this paper submission, 253 of these tests have already been merged into the projects (including 197 unit tests in the Luliyucoordinate-LeetCode project and 56 unit tests in the Rangerlee-HtmlParser project). Through running these tests, we identify 5 bugs, and 2 functional defects, and submitted corresponding issues to the project. The evaluation substantiates that MUT's test migration is both viable and beneficial across programming languages and different projects. © 2024 ACM.

关键词： Mapping

Transformer Doctor: Diagnosing and Treating Vision Transformers 38

学校读者我要写书评

暂无评论

Transformer Doctor: Diagnosing and Treating Vision Transform...

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Hu, Jiacong Chen, Hao Chen, Kejia Gao, Yang Ye, Jingwen Wang, Xingen Song, Mingli Feng, Zunlei College of Computer Science and Technology Zhejiang University China School of Software Technology Zhejiang University China National University of Singapore Singapore State Key Laboratory of Blockchain and Data Security Zhejiang University China Institute of Blockchain and Data Security China Bangsheng Technology Co. Ltd. China

Due to its powerful representational capabilities, Transformers have gradually become the mainstream model in the field of machine vision. However, the vast and complex parameters of Transformers impede researchers from gaining a deep understanding of their internal mechanisms, especially error mechanisms. Existing methods for interpreting Transformers mainly focus on understanding them from the perspectives of the importance of input tokens or internal modules, as well as the formation and meaning of features. In contrast, inspired by research on information integration mechanisms and conjunctive errors in the biological visual system, this paper conducts an in-depth exploration of the internal error mechanisms of Transformers. We first propose an information integration hypothesis for Transformers in the machine vision domain and provide substantial experimental evidence to support this hypothesis. This includes the dynamic integration of information among tokens and the static integration of information within tokens in Transformers, as well as the presence of conjunctive errors therein. Addressing these errors, we further propose heuristic dynamic integration constraint methods and rule-based static integration constraint methods to rectify errors and ultimately improve model performance. The entire methodology framework is termed as Transformer Doctor, designed for diagnosing and treating internal errors within transformers. Through a plethora of quantitative and qualitative experiments, it has been demonstrated that Transformer Doctor can effectively address internal errors in transformers, thereby enhancing model performance. For more information, please visit https://***/. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

Automating Comment Generation for Smart Contract from Bytecode

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Xiang, Jianhang Gao, Zhipeng Bao, Lingfeng Hu, Xing Chen, Jiayuan Xia, Xin The State Key Laboratory of Blockchain and Data Security Zhejiang University Hangzhou China The State Key Laboratory of Blockchain and Data Security Zhejiang University Shanghai China Blockchain and Data Security Research Institute Hangzhou China The Ohio State Univeristy Columbus United States

Recently, smart contracts have played a vital role in automatic financial and business transactions. To help end users without programming background to better understand the logic of smart contracts, previous studies have proposed models for automatically translating smart contract source code into their corresponding code summaries. However, in practice, only 13% of smart contracts deployed on the Ethereum blockchain are associated with source code. The practical usage of these existing tools is significantly restricted. Considering that bytecode is always necessary when deploying smart contracts, in this paper, we first introduce the task of automatically generating smart contract code summaries from bytecode. We propose a novel approach, named SmartBT (Smart contract Bytecode Translator) for automatically translating smart contract bytecode into fine-grained natural language description directly. Two key challenges are posed for this task: structural code logic hidden in bytecode and the huge semantic gap between bytecode and natural language descriptions. To address the first challenge, we transform bytecode into CFG (Control-Flow Graph) to learn code structural and logic details. Regarding the second challenge, we introduce an information retrieval component to fetch similar comments for filling the semantic gap. Then the structural input and semantic input are used to build an attentional sequence-to-sequence neural network model. The copy mechanism is employed to copy rare words directly from similar comments and the coverage mechanism is employed to eliminate repetitive outputs. The automatic evaluation results show that SmartBT outperforms a set of baselines by a large margin, and the human evaluation results show the effectiveness and potential of SmartBT in producing meaningful and accurate comments for smart contract code from bytecode directly. Copyright © 2025, The Authors. All rights reserved.

关键词： Semantics

SELF-EXPLAINED keyWORDS EMPOWER LARGE LANGUAGE MODELS FOR CODE GENERATION

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Fan, Lishui Chen, Mouxiang Liu, Zhongxin The State Key Laboratory of Blockchain and Data Security Zhejiang University China

Large language models (LLMs) have achieved impressive performance in code generation. However, due to the long-tail distribution of LLMs’ training data, low-frequency terms are typically underrepresented in the training process. Consequently, LLMs often misunderstand or overlook problem-specific, low-frequency keywords during code generation, compromising the accuracy of the generated code. To address this, we propose a novel technique named SEK (Self-Explained keywords), which empowers an LLM for better code generation by extracting and explaining the key terms in the problem description with the LLM itself and ranking them based on frequency. Comprehensive experiments across three benchmarks, i.e., HumanEval(+), MBPP(+), and APPS, with five representative LLMs, show that SEK can significantly improve LLMs in code generation, yielding substantial and consistent gains. For instance, SEK improves the Pass@1 of DeepSeek-Coder-V2-Instruct from 85.4% to 93.3% on the Humaneval benchmark. Further analysis confirms that SEK enables the LLMs to shift their attention from low-frequency keywords to their corresponding high-frequency counterparts. Copyright © 2024, The Authors. All rights reserved.

关键词： Codes (symbols)

Robust Watermarks Leak: Channel-Aware Feature Extraction Enables Adversarial Watermark Manipulation

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Ba, Zhongjie Zhang, Yitao Cheng, Peng Gong, Bin Zhang, Xinyu Wang, Qinglong Ren, Kui The State Key Laboratory of Blockchain and Data Security Zhejiang University Hangzhou China

Watermarking plays a key role in the provenance and detection of AI-generated content. While existing methods prioritize robustness against real-world distortions (e.g., JPEG compression and noise addition), we reveal a fundamental tradeoff: such robust watermarks inherently improve the redundancy of detectable patterns encoded into images, creating exploitable information leakage. To leverage this, we propose an attack framework that extracts leakage of watermark patterns through multi-channel feature learning using a pre-trained vision model. Unlike prior works requiring massive data or detector access, our method achieves both forgery and detection evasion with a single watermarked image. Extensive experiments demonstrate that our method achieves a 60% success rate gain in detection evasion and 51% improvement in forgery accuracy compared to state-of-the-art methods while maintaining visual fidelity. Our work exposes the robustness-stealthiness paradox: current "robust" watermarks sacrifice security for distortion resistance, providing insights for future watermark design. Copyright © 2025, The Authors. All rights reserved.

关键词： Information leakage

SWAT: ASystem-Wide Approach to Tunable Leakage Mitigation in Encrypted data Stores 50th

学校读者我要写书评

暂无评论

SWAT: ASystem-Wide Approach to Tunable Leakage Mitigation in...

50th International Conference on Very Large data Bases, VLDB 2024

作者： Zheng, Leqian Xu, Lei Wang, Cong Wang, Sheng Hu, Yuke Qin, Zhan Li, Feifei Ren, Kui City University of Hong Kong Hong Kong Nanjing University of Science and Technology China Alibaba Group China Zhejiang University The State Key Laboratory of Blockchain and Data Security China

Numerous studies have underscored the significant privacy risks associated with various leakage patterns in encrypted data stores. While many solutions have been proposed to mitigate these leak ages, they either (1) incur substantial overheads, (2) focus on specific subsets of leakage patterns, or (3) apply the same security notion across various workloads, thereby impeding the attainment of ne-tuned privacy efficiency trade-o s. In light of various detrimental leakage patterns, this paper starts with an investigation into which specific leakage patterns require our focus in the contexts of key-value, range-query, and dynamic workloads, respectively. Sub sequently, we introduce newsecurity notions tailored to the specfic privacy requirements of these workloads. Accordingly, we propose and instantiate Swat, an efficient construction that progressively enables these workloads, while provably mitigating system-wide leakage via a suite of algorithms with tunable privacy-efficiency trade-offs. We conducted extensive experiments and compiled a de tailed result analysis, showing the efficiency of our solution. Swat is about an order of magnitude slower than an encryption-only data store that reveals various leakage patterns and is two orders of magnitude faster than a trivial zero-leakage solution. Meanwhile, the performance of Swat remains highly competitive compared to other designs that mitigate specific types of leakage. © 2024, VLDB Endowment. All rights reserved.

关键词： Information leakage

Locally Private Estimator for Degree Distribution with Laplacian Mechanism 10

学校读者我要写书评

暂无评论

Locally Private Estimator for Degree Distribution with Lapla...

10th IEEE Smart World Congress, SWC 2024

作者： Li, Jiayu Hu, Yuke Li, Xiaoguang Zhou, Shiqi Wang, Yuxiang Li, Fenghua Niu, Ben Institute of Information Engineering Chinese Academy of Sciences China University of Chinese Academy of Sciences School of Cyber Security China Key Laboratory of Cyberspace Security Defense China Zhejiang University The State Key Laboratory of Blockchain and Data Security China Xidian University School of Cyber Engineering China

ISBN: (纸本)9798331520861

Local differential privacy (LDP) provides a solution such that an untrusted server can achieve privacy-preserving data analysis. Node degree is one of the most fundamental metrics in graph data analysis. Existing LDP-based methods can not balance degree distribution estimation and degree distribution characteristic estimation. In this paper, we propose a local private estimator for degree distribution with Laplacian mechanism (LEDL), which uses elastic net regression to estimate the degree distribution, selects the regularization weight to prevent overfitting or underfitting. We can aslo adopt the existing method to calculate degree distribution characteristics according to data with Laplacian noise. Experiments demonstrate the effectiveness of LEDL in degree distribution and Laplacian mechanism in degree distribution characteristic estimation, six metrics are applied across four datasets. © 2024 IEEE.

关键词： Differential privacy