- 天气预报中的对抗性观测Adversarial Observations in Weather Forecasting
Erik Imgrund, Thorsten Eisenhofer, Konrad Rieck会议页面 - 攻破纸质选票：投票遇上对抗机器学习Busting the Paper Ballot: Voting Meets Adversarial Machine Learning
Kaleel Mahmood, Caleb Manicke, Ethan Rathbun 等会议页面 - 评估生产级恶意软件检测系统对可迁移对抗攻击的鲁棒性Evaluating the robustness of a production malware detection system to transferable adversarial attacks
Milad Nasr, Yanick Fratantonio, Luca Invernizzi 等会议页面 - GASLITE：探索基于稠密嵌入的检索中的漏洞GASLITEing the Retrieval: Exploring Vulnerabilities in Dense Embedding-based Search
Matan Ben-Tov, Mahmood Sharif会议页面 - 一个代理骗所有模型：基于 CLIP 的通用、可迁移、定向对抗攻击One Surrogate to Fool Them All: Universal, Transferable, and Targeted Adversarial Attacks with CLIP
Binyan Xu, Xilin Dai, Di Tang 等会议页面
- 面向加壳可执行文件检测的对抗鲁棒汇编语言模型Adversarially Robust Assembly Language Model for Packed Executables Detection
非 AI 分会场Shijia Li, Jiang Ming, Lanqing Liu 等会议页面 - 像分析二进制一样分析 PDF：基于中间表示和语言模型的对抗鲁棒 PDF 恶意软件分析Analyzing PDFs like Binaries: Adversarially Robust PDF Malware Analysis via Intermediate Representation and Language Model
非 AI 分会场Side Liu, Jiang Ming, Guodong Zhou 等会议页面 - 从不完美标注的恶意软件数据中进行深度学习Deep Learning from Imperfectly Labeled Malware Data
Fahad Alotaibi, Euan Goodbrand, Sergio Maffeis会议页面 - k 近邻的精确鲁棒性认证Exact Robustness Certification of k-Nearest Neighbors
Francesco Ranzato, Ahmad Shakeel, Marco Zanella会议页面 - 给 LLM 上安全课：通过知识注入保障检索增强代码生成的安全Give LLMs a Security Course: Securing Retrieval-Augmented Code Generation via Knowledge Injection
非 AI 分会场Bo Lin, Shangwen Wang, Yihao Qin 等会议页面 - 加固深度神经网络二进制以抵御逆向工程攻击Hardening Deep Neural Network Binaries against Reverse Engineering Attacks
非 AI 分会场Zheng Zhong, Ruoyu Wu, Junpeng Wan 等会议页面 - 通过原像合成与属性细化对深度神经网络缺陷进行可证明修复Provable Repair of Deep Neural Network Defects by Preimage Synthesis and Property Refinement
Jianan Ma, Jingyi Wang, Qi Xuan 等会议页面 - RAG-WM：面向大模型检索增强生成的高效黑盒水印方法RAG-WM: An Efficient Black-Box Watermarking Approach for Retrieval-Augmented Generation of Large Language Models
Peizhuo Lv, Mengjie Sun, Hao Wang 等会议页面 - AI 生成内容隐空间水印的移除攻击与防御Removal Attack and Defense on AI Generated Content Latent-based Watermarking
De Zhang Lee, Han Fang, Hanyi Wang 等会议页面 - 重新思考图像生成模型中的机器遗忘Rethinking Machine Unlearning in Image Generation Models
Renyang Liu, Wenjie Feng, Tianwei Zhang 等会议页面 - SafeGuider：面向文生图模型的鲁棒实用内容安全控制SafeGuider: Robust and Practical Content Safety Control for Text-to-Image Models
Peigui Qi, Kunsheng Tang, Wenbo Zhou 等会议页面 - Sylva：通过协同微调为预训练模型定制个性化对抗防御Sylva: Tailoring Personalized Adversarial Defense in Pre-trained Models via Collaborative Fine-tuning
Tianyu Qi, Lei Xue, Yufeng Zhan 等会议页面 - UnsafeBench：在真实与 AI 生成图像上评测图像安全分类器UnsafeBench: Benchmarking Image Safety Classifiers on Real-World and AI-Generated Images
非 AI 分会场Yiting Qu, Xinyue Shen, Yixin Wu 等会议页面 - YouthSafe：面向青少年的 LLM 安全基准与防护模型YouthSafe: A Youth-Centric Safety Benchmark and Safeguard Model for Large Language Models
非 AI 分会场Yaman Yu, Yiren Liu, Yuqi Zhang 等会议页面
- 实用且安全的拜占庭鲁棒聚合器A Practical and Secure Byzantine Robust Aggregator
De Zhang Lee, Aashish Kolluri, Prateek Saxena 等会议页面 - 对抗性偏见从注入到蒸馏在语言模型中的级联传播Cascading Adversarial Bias from Injection to Distillation in Language Models
Harsh Chaudhari, Jamie Hayes, Matthew Jagielski 等会议页面 - FilterFL：面向联邦学习的基于知识过滤的无数据后门防御FilterFL: Knowledge Filtering-based Data-Free Backdoor Defense for Federated Learning
Yanxin Yang, Ming Hu, Xiaofei Xie 等会议页面 - 横向联邦学习中的超参数与后门抵抗On Hyperparameters and Backdoor-Resistance in Horizontal Federated Learning
非 AI 分会场Simon Lachnit, Ghassan Karame会议页面 - 通过对抗性误标注毒化文生图 AI 模型的可行性研究On the Feasibility of Poisoning Text-to-Image AI Models via Adversarial Mislabeling
Stanley Wu, Ronik Bhaskar, Anna Yoo Jeong Ha 等会议页面 - PoisonSpot：通过细粒度训练溯源追踪精确定位 clean-label 后门PoisonSpot: Precise Spotting of Clean-Label Backdoors via Fine-Grained Training Provenance Tracking
Philemon Hailemariam, Birhanu Eshete会议页面 - PET 加固深度学习模型中的幽灵威胁：不可见的配置诱导攻击The Phantom Menace in PET-Hardened Deep Learning Models: Invisible Configuration-Induced Attacks
Yiteng Peng, Dongwei Xiao, Zhibo Liu 等会议页面 - 实现模型参数空间中的后门隐蔽性Towards Backdoor Stealthiness in Model Parameter Space
Xiaoyun Xu, Zhuoran Liu, Stefanos Koffas 等会议页面 - VillainNet：沿精度-延迟 Pareto 前沿针对 SuperNet 的定向投毒攻击VillainNet: Targeted Poisoning Attacks Against SuperNets Along the Accuracy-Latency Pareto Frontier
David Oygenblik, Abhinav Vemulapalli, Animesh Agrawal 等会议页面 - 谁在幕后操纵？通过概念偏移评估 AI 训练与推理中的完整性与归因What’s Pulling the Strings? Evaluating Integrity and Attribution in AI Training and Inference through Concept Shift
Jiamin Chang, Haoyang Li, Hammond Pearce 等会议页面
- 匿名性揭秘：审计深度学习模型数据使用的实用框架Anonymity Unveiled: A Practical Framework for Auditing Data Use in Deep Learning Models
非 AI 分会场Zitao Chen, Karthik Pattabiraman会议页面 - Armadillo：带输入验证的鲁棒单服务器联邦学习安全聚合Armadillo: Robust Single-Server Secure Aggregation for Federated Learning with Input Validation
非 AI 分会场Yiping Ma, Yue Guo, Harish Karthikeyan 等会议页面 - LLM 能保护个人健康信息吗？医疗领域的隐私攻击与防御Can Personal Health Information Be Secured in LLM? Privacy Attack and Defense in the Medical Domain
非 AI 分会场Yujin Kang, Eunsun Kim, Yoon-Sik Cho会议页面 - Co-Prime：面向 FPGA 上隐私保护机器学习的协同设计框架Co-Prime: A Co-design Framework for Privacy Preserving Machine Learning on FPGA
Shuo Xu, Jiming Xu, Pengfei Xue 等会议页面 - CryptGNN：实现图神经网络的安全推理CryptGNN: Enabling Secure Inference for Graph Neural Networks
Pritam Sen, Yao Ma, Cristian Borcea会议页面 - DCMI：针对检索增强生成的差分校准成员推断攻击DCMI: A Differential Calibration Membership Inference Attack Against Retrieval-Augmented Generation
非 AI 分会场Xinyu Gao, Xiangtao Meng, Yingkai Dong 等会议页面 - 基于差分的微调大模型专有数据提取Differentiation-Based Extraction of Proprietary Data from Fine-tuned LLMs
Zongjie Li, Daoyuan Wu, Shuai Wang 等会议页面 - DivTrackee 对 DynTracker：面向动态人脸识别策略的多样化反人脸识别DivTrackee versus DynTracker: Promoting Diversity in Anti-Facial Recognition against Dynamic FR Strategy
Wenshu Fan, Minxing Zhang, Hongwei Li 等会议页面 - DPImageBench：差分隐私图像合成统一基准DPImageBench: A Unified Benchmark for Differentially Private Image Synthesis
Chen GONG, Kecen Li, Zinan Lin 等会议页面 - 在联邦学习中利用稀疏化：安全、高效且差分隐私的实现Harnessing Sparsification in Federated Learning: A Secure, Efficient, and Differentially Private Realization
非 AI 分会场Shuangqing Xu, Yifeng Zheng, Zhongyun Hua会议页面 - 作为隐私工具的成员推断攻击：可靠性、差异性与集成Membership Inference Attacks as Privacy Tools: Reliability, Disparity and Ensemble
Zhiqi Wang, Chengyu Zhang, Yuetian Chen 等会议页面 - MoEcho：利用侧信道攻击破坏 Mixture-of-Experts 大模型的用户隐私MoEcho: Exploiting Side-Channel Attacks to Compromise User Privacy in Mixture-of-Experts LLMs
Ruyi Ding, Tianhong Xu, Xinyi Shen 等会议页面 - Mosformer：面向大型 Transformer 的恶意安全三方推理框架Mosformer: Maliciously Secure Three-Party Inference Framework for Large Transformers
Ke Cheng, Yuheng Xia, Anxiao Song 等会议页面 - PLRV-O：通过隐私损失随机变量优化推进差分隐私深度学习PLRV-O: Advancing Differentially Private Deep Learning via Privacy Loss Random Variable Optimization
Qin Yang, Nicholas Stout, Meisam Mohammady 等会议页面 - PreferCare：通过水印注入与验证保护 LLM 对齐中的偏好数据集版权PreferCare: Preference Dataset Copyright Protection in LLM Alignment by Watermark Injection and Verification
Jian Lou, Chenyang Zhang, Xiaoyu Zhang 等会议页面 - 针对分布式大语言模型推理框架的 Prompt 推断攻击Prompt Inference Attack on Distributed Large Language Model Inference Frameworks
Xinjian Luo, Ting Yu, Xiaokui Xiao会议页面 - Prototype Surgery：通过软标签调整神经原型以实现高效机器遗忘Prototype Surgery: Tailoring Neural Prototypes via Soft Labels for Efficient Machine Unlearning
Gaoyang Liu, Xijie Wang, Zixiong Wang 等会议页面 - Riddle Me This！针对检索增强生成的隐蔽成员推断Riddle Me This! Stealthy Membership Inference for Retrieval-Augmented Generation
非 AI 分会场Ali Naseh, Yuefeng Peng, Anshuman Suri 等会议页面 - 保护图神经网络免受拓扑推断攻击Safeguarding Graph Neural Networks against Topology Inference Attacks
Jie Fu, Yuan Hong, Zhili Chen 等会议页面 - SCOPE：扩展客户端后处理以实现高效的隐私保护模型推理SCOPE: Expanding Client-Side Post-Processing for Efficient Privacy-Preserving Model Inference
Shenchen Zhu, Kai Chen, Yue Zhao 等会议页面 - Split Unlearning：拆分学习中的机器遗忘Split Unlearning
Yanna Jiang, Guangsheng Yu, QIN WANG 等会议页面 - TensorShield：用 TEE 保护关键 DNN 张量以守护端侧推理TensorShield: Safeguarding On-Device Inference by Shielding Critical DNN Tensors with TEE
Tong Sun, Bowen Jiang, Hailong Lin 等会议页面 - THOR：基于同态加密的安全 Transformer 推理THOR: Secure Transformer Inference with Homomorphic Encryption
Jungho Moon, Dongwoo Yoo, Xiaoqian Jiang 等会议页面 - WPC：基于 RNS-CKKS 的 CNN 推理权重明文压缩WPC: Weight Plaintext Compression for CNN Inference based on RNS-CKKS
Guiming Shi, Yuchen Wei, Shengyu Fan 等会议页面 - 无物可窃：通过系统向量缓解 LLM 的提示词泄露You Can't Steal Nothing: Mitigating Prompt Leakages in LLMs via System Vectors
Bochuan Cao, Yuanpu Cao, Changjiang Li 等会议页面 - 高精度零知识 AI 推理Zero-Knowledge AI Inference with High Precision
非 AI 分会场Arman Riasi, Haodi Wang, Rouzbeh Behnia 等会议页面 - ZORRO：面向拆分学习的零知识鲁棒性与隐私ZORRO: Zero-Knowledge Robustness and Privacy for Split Learning
Nojan Sheybani, Alessandro Pegoraro, Jonathan Knauer 等会议页面
- JsDeObsBench：评测 LLM 的 JavaScript 去混淆能力JsDeObsBench: Measuring and Benchmarking LLMs for JavaScript Deobfuscation
Guoqiang Chen, Xin Jin, Zhiqiang Lin会议页面 - 评测并增强大语言模型解决 CTF 挑战的能力Measuring and Augmenting Large Language Models for Solving Capture-the-Flag Challenges
非 AI 分会场Zimo JI, Daoyuan Wu, Wenyuan Jiang 等会议页面 - OCR-APT：用子图异常检测与 LLM 从审计日志重构 APT 攻击故事OCR-APT: Reconstructing APT Stories from Audit Logs using Subgraph Anomaly Detection and LLMs
非 AI 分会场Ahmed Aly, Essam Mansour, Amr Youssef会议页面 - Oedipus：LLM 增强的推理类 CAPTCHA 求解器Oedipus: LLM-enchanced Reasoning CAPTCHA Solver
Gelei Deng, Haoran Ou, Yi Liu 等会议页面 - PromeFuzz：知识驱动、用大语言模型生成 fuzzing harnessPromeFuzz: A Knowledge-Driven Approach to Fuzzing Harness Generation with Large Language Models
非 AI 分会场Yuwei Liu, Junquan Deng, Xiangkun Jia 等会议页面