2026· IEEE Transactions on Audio, Speech, and Language Processing· Vol 34, pp. 4203-4215· 0 citations· 45 references
Abstract
Current research on Audio Deepfake Detection (ADD) mainly relies on centralized training paradigms, while paying limited attention to privacy preservation under data localization regulations or user-consent constraints. In real-world scenarios, centralized data sharing may expose sensitive user information and hinder the practical deployment of ADD systems. To address this challenge, this paper proposes FLADD, a federated learning-based framework for ADD that supports data-localized training while retaining competitive detection performance. FLADD adopts a dual-branch architecture combining Res2Net's multi-scale spectral feature extraction with Rawformer's temporal modeling of raw waveforms, capturing complementary cues for distinguishing bona fide and spoofed speech. To improve robustness under heterogeneous client data, we introduce a Multi-Factor Adaptive Weighting (MFAW) mechanism, which dynamically adjusts aggregation weights by jointly considering sample size and local model performance, alleviating the bias of conventional size-based aggregation. In addition, FLADD employs a dual-prototype strategy, where global prototypes are constructed separately for the Res2Net and Rawformer branches to guide local optimization under heterogeneous data distributions, enhancing detection capability. Comprehensive experiments on the ASVspoof 2019 LA and ASVspoof 2021 LA evaluation sets demonstrate strong and competitive detection performance, with best observed Equal Error Rates of 0.285% and 2.045%, respectively. FLADD keeps raw audio local and provides empirical analyses of speaker-, content-, and transmitted-channel leakage.
Human Activity Recognition (HAR) using wearable sensor data has gained significant attention due to its applications in healthcare, smart environments, and human-computer interaction. However, traditional centralized learning approaches suffer from privacy risks and limited scalability. To address these challenges, thi...
Kamal Narayan Budholia, Ankur Pandey· International journal of com...· 0 citations
This work empirically diagnose the root cause across six per-layer methods and three speech-LLM architectures and proposes a two-pool allocation that normalises encoder and LLM parameters into independent pools, and shows that joint $\ell_2$ sensitivity and the original $(\varepsilon,\delta)$-DP guarantee are unchanged...
Jordi Luque, Fernando López, Aleix Sant· 0 citations
Side-Channel Analysis (SCA) methods critically depend on access to known plaintext/ciphertext, an assumption violated in many practical scenarios such as OCB encryption mode and the EMV payment protocol. Blind SCA addresses this challenge by operating without plaintext/ciphertext, relying instead on a three-phase pipel...
Jin-Tong Yu, Yuxuan Wang, Pei Cao et al.· IACR Transactions on Cryptog...· 0 citations
Existing audio deepfake detection (ADD) methods frequently exhibit limited generalization, as fixed bottleneck architectures fail to adapt to the heterogeneous information densities inherent in diverse spoofing artifacts. Moreover, standard augmentation often leads to a pathological robustness-accuracy trade-off. In th...
Jian-Qiao Cui, Hao-Xiang Zhu, Han-Biao Meng et al.· 0 citations
Federated learning (FL) enables multiple devices to collaboratively train machine learning models without sharing raw data, making it well-suited for Internet of Things (IoT) applications. However, this approach is not fully secure, as the exchanged gradients can still leak sensitive information. Attacks such as Deep L...