Signals 4 · free daily AI digest

Component-Aware Differential Privacy for Federated Multilingual Speech-LLMs

Paper recorded by Signals 4 on 2026-09-10 in cs.CL. Abstract reproduced from arXiv; link to the original below.

Published 2026-09-10 on arXiv · recorded by Signals 4 on 2026-09-11

Category: cs.CL · 自然语言处理 · first seen 2026-09-11

Abstract

Per-layer differential privacy (DP) clipping improves gradient fidelity in federated learning by allocating per-matrix clipping budgets proportional to parameter count. We show that this recipe breaks for speech large language models (speech-LLMs), when the acoustic encoder and the language decoder differ by an order of magnitude in update norm. Single-pool per-layer methods suffer \emph{cross-com

Read on arXiv →

#75 most recent of 186 cs.CL papers we have recorded · ↑ newer: The widening evaluation gap in medical large language model research 2 · ↓ older: RAG-Safety-Bench: Reliable Evaluation of Retrieval-Augmented LLM Safet
Cite this page: Component-Aware Differential Privacy for Federated Multilingual Speech-LLMs: the #75 most recent of 186 cs.CL papers we have recorded (as of 2026-09-10). Source: Signals 4 (Signals API) — https://data.jiangzhang.ca/signals4/t/papers/component-aware-differential-privacy-for-federated-multilingual-speech-llms.html
Free to quote with attribution to “Signals 4 (Signals API)”. Machine-readable: papers.json
Related: More cs.CL papers · arXiv signals · All papers · Today in AI
Get 4 AI signals a day by email — free.
Subscribe free → See all plans →
Get 4 AI signals a day by email — free
All models · All repos · By company · Daily editions