SL
Songze Li
cs.CRcs.LGcs.ITmath.ITcs.DCcs.CVcs.AIcs.CLstat.MLAdversarial training
On Valency
published · living versionsW_aku6bk7y·v1 · currentpublished
When Coding Agents Outpace Review: Evaluating PRD-Driven Development in Mature Enterprise Repositories
with Tianqi Liu, Yandong Guan, Wenkai Wang, Gang Yu +1
1 version
Preprints & journals
89 papers in the corpus · 2015–2026When Coding Agents Outpace Review: Evaluating PRD-Driven Development in Mature Enterprise Repositories2610.26394v1 · Tianqi Liu, Yandong Guan, Songze Li et al.2026 · 0 citationsarXivon Valency
Purifying Backdoored Large Vision-Language Models by Removing Hijacked Directions2610.09941v1 · Bojun Yang, Haochen Zhou, Zhifang Zhang et al.2026 · 0 citationsarXiv
Preserving Structure, Strengthening Signals: A Grammar-Driven Approach to Optimize the Trade-off Between Quality and Detectability in Code Watermarking2607.10210v2 · Licheng Yu, Aiwei Liu, Songze Li2026 · 0 citationsarXiv
When the Aggregator Cheats: Data-Free Backdoors in Federated LLM-based QA Systems2606.27511v1 · Chenqing Zhu, Yanbo Dai, Yulong Tian et al.2026 · 0 citationsarXiv
OptChain: Achieving Optimal Throughput of Permissionless Blockchains2606.23405v1 · Chunjiang Che, Songze Li, Xuechao Wang2026 · 0 citationsarXiv
OCELOT: Inference-Leakage Budgets for Privacy-Preserving LLM Agents2606.12341v1 · Jin Xie, Songze Li2026 · 0 citationsarXiv
PrivacyPeek: Auditing What LLM-Based Agents Acquire, Not Just What They Say2606.00152v2 · Mingxuan Zhang, Jiahui Han, Dadi Guo et al.2026 · 0 citationsarXiv
When Efficiency Backfires: Cascading LLMs Trigger Cascade Failure under Adversarial Attack2605.17288v1 · Zehan Sun, Dingfan Chen, Songze Li2026 · 0 citationsarXiv
BiAxisBias: Evaluating LLM Bias Beyond a Single Prompt and a Single Explanation2605.09041v2 · Jialing Gan, Junhao Dong, Songze Li2026 · 0 citationsarXiv
Cross-Modal Backdoors in Multimodal Large Language Models2605.07490v1 · Runhe Wang, Li Bai, Haibo Hu et al.2026 · 0 citationsarXiv
Hidden Ads: Behavior Triggered Semantic Backdoors for Advertisement Injection in Vision Language Models2603.27522v1 · Duanyi Yao, Changyue Li, Zhicong Huang et al.2026 · 0 citationsarXiv
Unveiling the Security Risks of Federated Learning in the Wild: From Research to Practice2603.20615v1 · Jiahao Chen, Zhiming Zhao, Yuwen Pu et al.2026 · 0 citationsarXiv
SEAL-Tag: Self-Tag Evidence Aggregation with Probabilistic Circuits for PII-Safe Retrieval-Augmented Generation2603.17292v1 · Jin Xie, Songze Li, Guang Cheng2026 · 1 citationarXiv
CIBER: A Comprehensive Benchmark for Security Evaluation of Code Interpreter Agents2602.19547v1 · Lei Ba, Qinbin Li, Songze Li2026 · 0 citationsarXiv
Noise as a Probe: Membership Inference Attacks on Diffusion Models Leveraging Initial Noise2601.21628v2 · Puwei Lian, Yujun Cai, Songze Li et al.2026 · 0 citationsarXiv
Attributing and Exploiting Safety Vectors through Global Optimization in Large Language Models2601.15801v1 · Fengheng Chu, Jiahao Chen, Yuhong Wang et al.2026 · 0 citationsarXiv
CODE: A Contradiction-Based Deliberation Extension Framework for Overthinking Attacks on Retrieval-Augmented Generation2601.13112v1 · Xiaolei Zhang, Xiaojun Jia, Liquan Chen et al.2026 · 0 citationsarXiv
MASH: Evading Black-Box AI-Generated Text Detectors via Style Humanization2601.08564v2 · Yongtong Gu, Songze Li, Xia Hu2026 · 0 citationsarXiv
Knowledge-Driven Multi-Turn Jailbreaking on Large Language Models2601.05445v1 · Songze Li, Ruishi He, Xiaojun Jia et al.2026 · 0 citationsarXiv
Tackling Resource-Constrained and Data-Heterogeneity in Federated Learning with Double-Weight Sparse Pack2601.01840v1 · Qiantao Yang, Liquan Chen, Mingfu Xue et al.2026 · 2 citationsarXiv
Odysseus: Jailbreaking Commercial Multimodal LLM-integrated Systems via Dual Steganography2512.20168v1 · Songze Li, Jiameng Cheng, Yiming Li et al.2025 · 0 citationsarXiv
Bi-Erasing: A Bidirectional Framework for Concept Removal in Diffusion Models2512.13039v2 · Hao Chen, Yiwei Wang, Songze Li2025 · 0 citationsarXiv
FuncPoison: Poisoning Function Library to Hijack Multi-agent Autonomous Driving Systems2509.24408v2 · Yuzhen Long, Songze Li2025 · 0 citationsarXiv
Visual CoT Makes VLMs Smarter but More Fragile2509.23789v1 · Chunxue Xu, Yiwei Wang, Yujun Cai et al.2025 · 0 citationsarXiv
Responsible Diffusion: A Comprehensive Survey on Safety, Ethics, and Trust in Diffusion Models2509.22723v1 · Kang Wei, Xin Yuan, Fushuo Huo et al.2025 · 0 citationsarXiv
SA3C-ID: a novel network intrusion detection model using feature selection and adversarial training.40989373 · Huang, Wanwei, Tian, Haobin, Wang, Lei et al.2025 · 4 citationsPeerJ. Computer science. 2025;11:e3089
Asymptotically Optimal Secure Aggregation for Wireless Federated Learning with Multiple Servers2506.23680v1 · Zhenhao Huang, Kai Liang, Yuanming Shi et al.2025 · 0 citationsarXiv
LLMs Cannot Reliably Judge (Yet?): A Comprehensive Assessment on the Robustness of LLM-as-a-Judge2506.09443v3 · Songze Li, Chuokun Xu, Jiaying Wang et al.2025 · 0 citationsarXiv
TooBadRL: Trigger Optimization to Boost Effectiveness of Backdoor Attacks on Deep Reinforcement Learning2506.09562v3 · Mingxuan Zhang, Oubo Ma, Kang Wei et al.2025 · 0 citationsarXiv
Privacy-preserving Prompt Personalization in Federated Learning for Multimodal Large Language Models2505.22447v1 · Sizai Hou, Songze Li, Baturalp Buyukates2025 · 0 citationsarXiv
Enhancing Membership Inference Attacks on Diffusion Models from a Frequency-Domain Perspective2505.20955v5 · Puwei Lian, Yujun Cai, Songze Li et al.2025 · 0 citationsarXiv
SecurePay: Enabling Secure and Fast Payment Processing for Platform Economy2505.17466v1 · Junru Lin, Mingzhe Liu, Songze Li et al.2025 · 3 citationsarXiv
ReCIT: Reconstructing Full Private Data from Gradient in Parameter-Efficient Fine-Tuning of Large Language Models2504.20570v1 · Jin Xie, Ruishi He, Songze Li et al.2025 · 0 citationsarXiv
TrojanDam: Detection-Free Backdoor Defense in Federated Learning through Proactive Model Robustification utilizing OOD Data2504.15674v1 · Yanbo Dai, Songze Li, Zihan Gan et al.2025 · 0 citationsarXiv
How does Watermarking Affect Visual Language Models in Document Understanding?2504.01048v2 · Chunxue Xu, Yiwei Wang, Bryan Hooi et al.2025 · 0 citationsarXiv
When Data Manipulation Meets Attack Goals: An In-depth Survey of Attacks for VLMs2502.06390v2 · Aobotao Dai, Xinyu Ma, Lei Chen et al.2025 · 0 citationsarXiv
Arbitrary-Threshold Fully Homomorphic Encryption with Lower Complexity2501.11235v1 · Yijia Chang, Songze Li2025 · 0 citationsarXiv
Rethinking Knowledge in Distillation: An In-context Sample Retrieval Perspective2501.07040v1 · Jinjing Zhu, Songze Li, Lin Wang2025 · 0 citationsarXiv
Hidden Data Privacy Breaches in Federated Learning2411.18269v1 · Xueluan Gong, Yuji Wang, Shuaike Li et al.2024 · 0 citationsarXiv
DeDe: Detecting Backdoor Samples for SSL Encoders via Decoders2411.16154v2 · Sizai Hou, Songze Li, Duanyi Yao2024 · 2 citationsarXiv
Navigating the Risks: A Survey of Security, Privacy, and Ethics Threats in LLM-Based Agents2411.09523v1 · Yuyou Gan, Yong Yang, Zhe Ma et al.2024 · 8 citationsarXiv
Agent-Oriented Planning in Multi-Agent Systems2410.02189v2 · Ao Li, Yuexiang Xie, Songze Li et al.2024 · 3 citationsarXiv
Constructing Adversarial Examples for Vertical Federated Learning: Optimal Client Corruption through Multi-Armed Bandit2408.04310v1 · Duanyi Yao, Songze Li, Ye Xue et al.2024 · 2 citationsarXiv
Manifoldchain: Maximizing Blockchain Throughput via Bandwidth-Clustered Sharding2407.16295v1 · Chunjiang Che, Songze Li, Xuechao Wang2024 · 8 citationsarXiv
PriRoAgg: Achieving Robust Model Aggregation with Minimum Privacy Leakage for Federated Learning2407.08954v2 · Sizai Hou, Songze Li, Tayyebeh Jahani-Nezhad et al.2024 · 12 citationsarXiv
MSfusion: A Dynamic Model Splitting Approach for Resource-Constrained Machines to Collaboratively Train Larger Models2407.03622v1 · Jin Xie, Songze Li2024 · 0 citationsarXiv
BackdoorIndicator: Leveraging OOD Data for Proactive Backdoor Detection in Federated Learning2405.20862v1 · Songze Li, Yanbo Dai2024 · 2 citationsarXiv
Towards Client Driven Federated Learning2405.15407v1 · Songze Li, Chenqing Zhu2024 · 1 citationarXiv
TUNI: A Textual Unimodal Detector for Identity Inference in CLIP Models2405.14517v2 · Songze Li, Ruoxi Cheng, Xiaojun Jia2024 · 1 citationarXiv
URVFL: Undetectable Data Reconstruction Attack on Vertical Federated Learning2404.19582v2 · Duanyi Yao, Songze Li, Xueluan Gong et al.2024 · 4 citationsarXiv
Career total: 132 works. 89 are in this corpus.Showing the 50 most recent.
Profile built from the corpus for this byline.
Author records are still filling in while the Hub is in alpha. If this is your page, you'll be able to claim it soon. Spot a mistake? Tell us.