·
DataBubble
  • Home
  • Models
  • News
  • Compare
  • Boards
  • Pricing
  • About
  • Newsletter
  • Methodology
  • Contact
Latest
Retrieval Augmented Generation Compositionality7h◆VeriEvolve: Evaluator-First Evolution for Tool-Using Agents7h◆TNODEV: Toolbox for Neural ODE Verification7h◆Learning to Interfere in Non-Orthogonal Multiple-Access Joint Source-Channel Coding7h◆On the Expressive Power and Limitations of Multi-Layer SSMs7h◆From Weather to Enhanced Weathering: Weather Foundation Models for Carbon Dioxide Removal at Operational Scale7h◆Predictor-Guided Latent Space Codon Optimization for Maximizing Protein Expression7h◆A Generalisation Signal Need Not Be a Model-Selection Signal7h◆Mitigating Reward Hacking in Realistic Scenarios Through Representation Engineering7h◆EgoHand: Monocular 3D Hand Reconstruction from Egocentric Video7h◆Learning to Incentivise: Using Reinforcement Learning for Sustainable Urban Mobility7h◆ChemBART: Pre-training a Reaction Language Model for Comprehensive Synthesis Planning7h◆Kernel-Managed Shared Memory for System-Wide Personalization7h◆Uber rival inDrive eyes 2027 robotaxi tests in Central Asia with Pony.ai and Nvidia7h◆Automating ML for Science: Can Frontier Agents Climb Scientific Hills in the Wild?7h◆AnnoCaseLaw: Legal Annotation Benchmark of Civil Negligence7h◆Improved Training of Strong Stochastic Flow Maps7h◆Think Harder Where It Matters: Item-Level Adaptive Recursion for Sequential Recommendation7h◆HyperDrone: Simulation-to-Reality Transfer at Millions of Frames per Second7h◆Structured Temporal Causality for Interpretable Multivariate Time Series Anomaly Detection7h◆Retrieval Augmented Generation Compositionality7h◆VeriEvolve: Evaluator-First Evolution for Tool-Using Agents7h◆TNODEV: Toolbox for Neural ODE Verification7h◆Learning to Interfere in Non-Orthogonal Multiple-Access Joint Source-Channel Coding7h◆On the Expressive Power and Limitations of Multi-Layer SSMs7h◆From Weather to Enhanced Weathering: Weather Foundation Models for Carbon Dioxide Removal at Operational Scale7h◆Predictor-Guided Latent Space Codon Optimization for Maximizing Protein Expression7h◆A Generalisation Signal Need Not Be a Model-Selection Signal7h◆Mitigating Reward Hacking in Realistic Scenarios Through Representation Engineering7h◆EgoHand: Monocular 3D Hand Reconstruction from Egocentric Video7h◆Learning to Incentivise: Using Reinforcement Learning for Sustainable Urban Mobility7h◆ChemBART: Pre-training a Reaction Language Model for Comprehensive Synthesis Planning7h◆Kernel-Managed Shared Memory for System-Wide Personalization7h◆Uber rival inDrive eyes 2027 robotaxi tests in Central Asia with Pony.ai and Nvidia7h◆Automating ML for Science: Can Frontier Agents Climb Scientific Hills in the Wild?7h◆AnnoCaseLaw: Legal Annotation Benchmark of Civil Negligence7h◆Improved Training of Strong Stochastic Flow Maps7h◆Think Harder Where It Matters: Item-Level Adaptive Recursion for Sequential Recommendation7h◆HyperDrone: Simulation-to-Reality Transfer at Millions of Frames per Second7h◆Structured Temporal Causality for Interpretable Multivariate Time Series Anomaly Detection7h◆
News/When Helpful Context Leaks: Privacy Risks in Domain-Adapted ASR
arxiv
PublishedMay 28, 2026 at 4:00 AM
▼bearish

When Helpful Context Leaks: Privacy Risks in Domain-Adapted ASR

Source
arxiv.orgfull article ↗
Read on arxiv→
Publisher summary· verbatim

arXiv:2605.28211v1 Announce Type: new Abstract: SpeechLLMs are increasingly deployed in professional settings where domain customisation is standard practice: users supply context in prompts with sensitive information, fine-tune on proprietary recordings, or both. We identify and systematically inve

Stay posted· Newsletter

Three short emails a week: Monday, Wednesday and Friday. Top movers, new models and the most-referenced story.

// no spam · unsubscribe one-click · free forever

Your email is used only to send The Bubble Brief. Privacy policy

Discussion
Source
↗
arxiv
Read original ↗All from arxiv →
Tags
04
#privacy#customisation#speech#leakage

No replies yet. Be first.

Source
↗
arxiv
Read original ↗All from arxiv →
Tags
04
#privacy#customisation#speech#leakage

Related coverage

More from ARXIV
arxivRetrieval Augmented Generation Compositionality7harxivVeriEvolve: Evaluator-First Evolution for Tool-Using Agents7harxivTNODEV: Toolbox for Neural ODE Verification7harxivLearning to Interfere in Non-Orthogonal Multiple-Access Joint Source-Channel Coding7h
The Bubble Brief
MON · WED · FRI

Read privacy insights every Monday, Wednesday and Friday — top movers, new models, story of the week.

// no spam · unsubscribe one-click · free forever

Your email is used only to send The Bubble Brief. Privacy policy

Originally published on arxiv ↗
PrivacyTermsRefundsLegal notice|AboutMethodologyContact

© 2026 Databubble

Built by Marouane Gazouzi · Readyo, AI interview coach

HomeModelsNews