DP-OPD: Differentially Private On-Policy Distillation for Language Models

Large language models (LLMs) are increasingly adapted to proprietary and domain-specific corpora that contain sensitive information, creating a tension between formal privacy guarantees and efficient deployment through model compression. Differential privacy (DP), typically enforced via DP-SGD, provides record-level protection but often incurs substantial utility loss in autoregressive generation, where optimization noise can amplify exposure bias and compounding errors along long rollouts. Existing approaches to private distillation either apply DP-SGD to both teacher and student, worsening computation and the privacy--utility tradeoff, or rely on DP synthetic text generation from a DP-trained teacher, avoiding DP on the student at the cost of DP-optimizing a large teacher and introducing an offline generation pipeline. We propose \textbf{Differentially Private On-Policy Distillation (DP-OPD)}, a synthesis-free framework that enforces privacy solely through DP-SGD on the student while leveraging a frozen teacher to provide dense token-level targets on \emph{student-generated} trajectories. DP-OPD instantiates this idea via \emph{private generalized knowledge distillation} on continuation tokens. Under a strict privacy budget ($\varepsilon=2.0$), DP-OPD improves perplexity over DP fine-tuning and off-policy DP distillation, and outperforms synthesis-based DP distillation (Yelp: 44.15$\rightarrow$41.68; BigPatent: 32.43$\rightarrow$30.63), while substantially simplifying the training pipeline. In particular, \textbf{DP-OPD collapses private compression into a single DP student-training loop} by eliminating DP teacher training and offline synthetic text generation. Code will be released upon publication at https://github.com/khademfatemeh/dp_opd.

ECGBERT: Understanding Hidden Language of ECGs with Self-Supervised Representation Learning

In the medical field, current ECG signal analysis approaches rely on supervised deep neural networks trained for specific tasks that require substantial amounts of labeled data. However, our paper introduces ECGBERT, a self-supervised representation learning approach that unlocks the underlying language of ECGs. By unsupervised pre-training of the model, we mitigate challenges posed by the lack of well-labeled and curated medical data. ECGBERT, inspired by advances in the area of natural language processing and large language models, can be fine-tuned with minimal additional layers for various ECG-based problems. Through four tasks, including Atrial Fibrillation arrhythmia detection, heartbeat classification, sleep apnea detection, and user authentication, we demonstrate ECGBERT's potential to achieve state-of-the-art results on a wide variety of tasks.