The paper introduces S2-WEF, a dynamic free‑rider detection method for cross‑silo federated learning. It simulates weight‑evolving‑frequency (WEF) patterns of global‑model‑mimicking attacks on the server side and compares them with clients’ submitted WEFs, combining this similarity with a deviation score from mutual WEF comparisons. Using two‑dimensional clustering and per‑score classification, S2‑WEF can identify clients that switch to free‑riding during training, achieving robust detection across four datasets and five attack types without requiring proxy datasets or pre‑training.
arXiv:2604. 04611v2 Announce Type: replace Abstract: Federated learning (FL) enables multiple clients to collaboratively train a global model by aggregating local updates without sharing private data.
arXiv:2601. 07177v5 Announce Type: replace-cross Abstract: Federated learning (FL) addresses privacy and data-silo issues in the training of large language models (LLMs).
arXiv:2607. 07314v1 Announce Type: cross Abstract: Federated learning (FL) avoids explicit data exposure by keeping raw data on local clients, yet privacy risks remain in the training process and the learned model itself.
arXiv:2509. 11974v2 Announce Type: replace-cross Abstract: Federated Learning (FL) enables collaborative model training among clients without centralising data, making it a widely adopted privacy-enhancing technology (PET).
The paper introduces Fed-ADR, a coordinated attack framework where a malicious orchestrator server directs heterogeneous adversarial clients to adapt their gradient updates in real time, thereby evading existing federated learning defenses and drastically reducing global model accuracy. It also presents a lightweight detection mechanism that estimates true client gradients from historical data to spot coordinated attacks, and an in-situ recovery method that restores model performance without restarting training. Experiments on MNIST, Fashion‑MNIST, and CIFAR‑10 show the attack can drop accuracy from over 90% to below 10%, while the defense can recover accuracy to above 90% within a few rounds at a computational cost at least 20× lower than retraining from scratch.
arXiv:2606. 07857v1 Announce Type: cross Abstract: The rise of edge-based machine learning has enabled distributed adaptation of language models across mobile and IoT devices, offering privacy preservation and real-time responsiveness.
FedLNS is a server‑side framework that screens federated learning updates by representing each client’s contribution through changes in trainable normalization‑layer parameters, creating lightweight signatures that can be compared against a history‑aware cross‑client reference. The method requires no extra client‑to‑server communication, raw data, or labeled attack examples, and after screening, the remaining full‑model updates are aggregated with standard federated learning rules. Experiments on GPT‑style, BERT‑style, and LLaMA‑style models with 200 clients demonstrate that FedLNS achieves lower test perplexity than six baselines even when 40% of the population performs target manipulation under both IID and non‑IID data partitions.
Federated learning (FL) avoids explicit data exposure by keeping raw data on local clients, yet privacy risks remain in the training process and the learned model itself. Recently, centralized Taking Away Training Data (TATD) attacks have shown that malicious training could abuse the memorization capacity of deep models to store and later recover training data.
arXiv:2608. 04073v1 Announce Type: cross Abstract: Personalized Federated Learning (PFL) has emerged as a promising solution for intrusion detection in heterogeneous IoT environments, as it can improve local adaptation under highly Non-Independent and Identically Distributed (non-IID) data distributions.
The paper introduces FAB, an attack that uses meta‑learning to embed dormant adversarial behaviors into large language models (LLMs). These behaviors remain inactive until the model is finetuned by downstream users, at which point the model can exhibit unwanted actions such as unsolicited advertising, jailbreakability, or over‑refusal. FAB is shown to be effective across multiple LLMs and resilient to various finetuning settings.
The paper introduces D-ADD, a plug‑in defense for image classification models that protects against model‑stealing attacks. It uses a non‑parametric detector called Account‑aware Distribution Discrepancy (ADD) to identify malicious queries by modeling each class as a multivariate normal distribution and computing weighted distribution discrepancies. With an enhanced version ADD$^+$ that handles domain shifts and combined with random‑based prediction poisoning, D-ADD offers strong protection while minimally affecting benign users in both soft‑ and hard‑label settings.
arXiv:2606. 19023v1 Announce Type: cross Abstract: The growing reliance on pre-trained Machine Learning (ML) models has introduced new attack surfaces.