Entity · paper

Learning to Hear Hesitation: Continual Learning for Disfluency-Aware ASR

paperactivelearning-to-hear-hesitation-continual-learning-for-disfluency-aware-asr-3cd7bf2f·1 events·first seen Jun 15, 2026

Aliases: Learning to Hear Hesitation: Continual Learning for Disfluency-Aware ASR

More like this (12)

Progressive Refinement: An Iterative Pseudo-Labeling Approach for Mandarin-English Code-Switching ASR Efficient ASR Training with Conversations that Never Happened Audio-Native Speech Recognition with a Frozen Discrete-Diffusion Language Model SpeechLLM Meets Federated Learning for End-to-End ASR: English and Italian Case Studies FlowEdit: Associative Memory for Lifelong Pronunciation Adaptation in Flow-Matching TTS On the Limits of Prompt-Conditioned Language Models as General-Purpose Learners Real-Time Voice AI Hears but Does Not Listen Interleaved Speech Language Models Latently Work In Text Leveraging Audio-LLMs to Filter Speech-to-Speech Training Data Graded Entity-Familiarity Readouts in Language Models: Polish Adaptation, Cross-Language Robustness, and Refusal Steering REDDIT: Correcting Model-Generated Timestamp Drift in ASR without Forgetting via Replay-Based Distribution Editing Adaptive Multi-Step Lookahead Decoding for Diffusion Language Models

Recent events (1)

3arXiv · cs.CL·Jun 15, 2026·source ↗

Continual learning approach for disfluency-aware ASR with explicit disfluency tokens

A new arXiv preprint addresses the challenge of transcribing disfluent speech (hesitations, repetitions, fillers) in ASR systems, which typically omit such markers causing information loss. The authors introduce explicit disfluency tokens into a pretrained ASR model and apply continual learning to adapt across datasets with varying disfluency distributions while mitigating catastrophic forgetting. The work identifies a trade-off between disfluency marker learning and general ASR performance, and finds a consistent cross-attention head mechanism shared across continual learning methods.

Learning to Hear Hesitation: Continual Learning for Disfluency-Aware ASR