Computer Security – ESORICS 2025

.- Time-Distributed Backdoor Attacks on Federated Spiking Learning. .- TATA: Benchmark NIDS Test Sets Assessment and Targeted Augmentation. .- Abuse-Resistant Evaluation of AI-as-a-Service via Function-Hiding Homomorphic Signatures. .- PriSM: A Privacy-friendly Support vector Machine. .- Towards Context-Aware Log Anomaly Detection Using Fine-Tuned Large Language Models. .- PROTEAN: Federated Intrusion Detection in Non-IID Environments through Prototype-Based Knowledge Sharing. .- KeTS: Kernel-based Trust Segmentation against Model Poisoning Attacks. .- Machine Learning Vulnerabilities in 6G: Adversarial Attacks and Their Impact on Channel Gain Prediction and Resource Allocation in UC-CF-mMIMO. .- FuncVul: An Effective Function Level Vulnerability Detection Model using LLM and Code Chunk. .- LUMIA: Linear probing for Unimodal and MultiModal Membership Inference Attacks leveraging internal LLM states. .- Membership Privacy Evaluation in Deep Spiking Neural Networks. .- DUMB and DUMBer: Is Adversarial Training Worth It in the Real World?. .- Countering Jailbreak Attacks with Two-Axis Pre-Detection and Conditional Warning Wrappers. .- How Dataset Diversity Affects Generalization in ML-based NIDS. .- Llama-based source code vulnerability detection: Prompt engineering vs Finetuning.  .- DBBA: Diffusion-based Backdoor Attacks on Open-set Face Recognition Models. .- Evaluation of Autonomous Intrusion Response Agents In Adversarial and Normal Scenarios. .- Trigger-Based Fragile Model Watermarking for Image Transformation Networks. .- Let the Noise Speak: Harnessing Noise for a Unified Defense Against Adversarial and Backdoor Attacks. .- On the Adversarial Robustness of Graph Neural Networks with Graph Reduction. .- SecureT2I: No More Unauthorized Manipulation on AI Generated Images from Prompts. .- GANSec: Enhancing Supervised Wireless Anomaly Detection Robustness through Tailored Conditional GAN Augmentation. .- Fine-Grained Data Poisoning Attack to Local Differential Privacy Protocols for Key-Value Data. .- The DCR Delusion: Measuring the Privacy Risk of Synthetic Data. .- StructTransform: A Scalable Attack Surface for Safety-Aligned Large Language Models.