English
Related papers

Related papers: Private Speech Classification without Collapse: St…

200 papers

Differential Privacy (DP) is becoming central to large-scale training as privacy regulations tighten. We revisit how DP noise interacts with adaptivity in optimization through the lens of stochastic differential equations, providing the…

Deep learning techniques based on neural networks have shown significant success in a wide range of AI tasks. Large-scale training datasets are one of the critical factors for their success. However, when the training datasets are…

Cryptography and Security · Computer Science 2019-12-23 Lei Yu , Ling Liu , Calton Pu , Mehmet Emre Gursoy , Stacey Truex

On-policy self-distillation (SD) improves LLM reasoning by using teacher-side privileged information (PI) to turn sparse verifier outcomes into dense token-level supervision. Existing methods usually assume trusted PI, such as reference…

Computation and Language · Computer Science 2026-05-28 Jiazhen Huang , Xiao Chen , Xiao Luo , Yong Dai , Senkang Hu , Yuzhi Zhao

On-policy distillation (OPD) and on-policy self-distillation (OPSD) have emerged as promising post-training methods for large language models, offering dense token-level supervision on trajectories sampled from the model's own policy.…

Artificial Intelligence · Computer Science 2026-05-26 Siqi Zhu , Xuyan Ye , Hongyu Lu , Weiye Shi , Ge Liu

We study gradient descent under linearly correlated noise. Our work is motivated by recent practical methods for optimization with differential privacy (DP), such as DP-FTRL, which achieve strong performance in settings where privacy…

Machine Learning · Computer Science 2024-01-17 Anastasia Koloskova , Ryan McKenna , Zachary Charles , Keith Rush , Brendan McMahan

The hidden state threat model of differential privacy (DP) assumes that the adversary has access only to the final trained machine learning (ML) model, without seeing intermediate states during training. However, the current privacy…

Machine Learning · Computer Science 2025-05-27 Rob Romijnders , Antti Koskela

Models need to be trained with privacy-preserving learning algorithms to prevent leakage of possibly sensitive information contained in their training data. However, canonical algorithms like differentially private stochastic gradient…

Machine Learning · Computer Science 2022-10-06 Yannis Cattan , Christopher A. Choquette-Choo , Nicolas Papernot , Abhradeep Thakurta

Training large neural networks with meaningful/usable differential privacy security guarantees is a demanding challenge. In this paper, we tackle this problem by revisiting the two key operations in Differentially Private Stochastic…

Machine Learning · Computer Science 2022-10-11 Hanshen Xiao , Jun Wan , Srinivas Devadas

Protecting privacy in learning while maintaining the model performance has become increasingly critical in many applications that involve sensitive data. Private Gradient Descent (PGD) is a commonly used private learning framework, which…

Machine Learning · Computer Science 2022-10-20 Junyuan Hong , Zhangyang Wang , Jiayu Zhou

In this work we analyze the sample complexity of classification by differentially private algorithms. Differential privacy is a strong and well-studied notion of privacy introduced by Dwork et al. (2006) that ensures that the output of an…

Data Structures and Algorithms · Computer Science 2015-09-15 Vitaly Feldman , David Xiao

In this paper we tackle the challenge of making the stochastic coordinate descent algorithm differentially private. Compared to the classical gradient descent algorithm where updates operate on a single model vector and controlled noise…

Machine Learning · Computer Science 2021-03-16 Georgios Damaskinos , Celestine Mendler-Dünner , Rachid Guerraoui , Nikolaos Papandreou , Thomas Parnell

Diffusion models have shown strong performance in speech enhancement, but their real-time applicability has been limited by multi-step iterative sampling. Consistency distillation has recently emerged as a promising alternative by…

Audio and Speech Processing · Electrical Eng. & Systems 2026-05-19 Liang Xu , Longfei Felix Yan , W. Bastiaan Kleijn

The ever-growing learning model size nowadays challenges the communication efficiency and privacy preservation of the traditional federated learning (FL). In this paper, we propose a novel differentially private (DP) over-the-air federated…

Information Theory · Computer Science 2025-08-12 Zihao Hu , Jia Yan , Ying-Jun Angela Zhang

On-policy distillation offers dense, per-token supervision for training reasoning models; however, it remains unclear under which conditions this signal is beneficial and under which it is detrimental. Which teacher model should be used,…

A major challenge for machine learning is increasing the availability of data while respecting the privacy of individuals. Here we combine the provable privacy guarantees of the differential privacy framework with the flexibility of…

Machine Learning · Statistics 2019-01-18 Michael Thomas Smith , Max Zwiessele , Neil D. Lawrence

Differentially Private Stochastic Gradient Descent (DP-SGD) and its variants have been proposed to ensure rigorous privacy for fine-tuning large-scale pre-trained language models. However, they rely heavily on the Gaussian mechanism, which…

Cryptography and Security · Computer Science 2024-05-30 Qin Yang , Meisam Mohammad , Han Wang , Ali Payani , Ashish Kundu , Kai Shu , Yan Yan , Yuan Hong

While deep models have proved successful in learning rich knowledge from massive well-annotated data, they may pose a privacy leakage risk in practical deployment. It is necessary to find an effective trade-off between high utility and…

Machine Learning · Computer Science 2024-09-05 Shiming Ge , Bochao Liu , Pengju Wang , Yong Li , Dan Zeng

Certified machine unlearning can be achieved via noise injection leading to differential privacy guarantees, where noise is calibrated to worst-case sensitivity. Such conservative calibration often results in performance degradation,…

Machine Learning · Computer Science 2026-05-25 Hanna Benarroch , Jamal Atif , Olivier Cappé

Applications of Differential Privacy (DP) in NLP must distinguish between the syntactic level on which a proposed mechanism operates, often taking the form of $\textit{word-level}$ or $\textit{document-level}$ privatization. Recently,…

Computation and Language · Computer Science 2024-07-02 Stephen Meisenbacher , Maulik Chevli , Florian Matthes

Amplification by subsampling is one of the main primitives in machine learning with differential privacy (DP): Training a model on random batches instead of complete datasets results in stronger privacy. This is traditionally formalized via…

Cryptography and Security · Computer Science 2024-11-04 Jan Schuchardt , Mihail Stoian , Arthur Kosmala , Stephan Günnemann