L3-SE reduces linguistic hallucination in LM-based speech enhancement by distilling noise-invariant acoustic-semantic representations from noisy inputs to condition an autoregressive decoder-only language model.
SenSE: Semantic-Aware High-Fidelity Universal Speech Enhancement
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
abstract
Generative Universal Speech Enhancement (USE) methods aim to leverage generative models to improve speech quality under various types of distortions. However, existing generative speech enhancement methods often suffer from semantic inconsistency in the generated outputs. Therefore, we propose SenSE, a novel two-stage generative universal speech enhancement framework, by modeling semantic priors with a language model, the flow matching-based speech enhancement process is guided to generate semantically faithful speech, thereby effectively improving context fidelity. In addition, we introduce a dual-path masked conditioning training strategy that enables flow matching-based enhancement to flexibly integrate multi-source conditioning signals from degraded speech, semantic tokens, and reference speech, thereby improving model flexibility and adaptability. Experimental results demonstrate that SenSE achieves state-of-the-art performance among generative speech enhancement models and exhibits a high performance ceiling, particularly under challenging distortion conditions. Codes and demos are available at https://github.com/ASLP-lab/SenSE.
citation-role summary
method 1
citation-polarity summary
fields
eess.AS 3years
2026 3roles
method 1polarities
use method 1representative citing papers
UniPASE extends the low-hallucination PASE framework to universal speech enhancement, restoring seven distortion types at flexible sampling rates with better word-error and speaker-similarity scores than prior generative systems.
PhASE-Flow performs phonetic-conditioned acoustic flow matching entirely in SSL representation space for speech enhancement and reports competitive perceptual quality with only four sampling steps.
citing papers explorer
-
Reducing Linguistic Hallucination in LM-Based Speech Enhancement via Noise-Invariant Acoustic-Semantic Distillation
L3-SE reduces linguistic hallucination in LM-based speech enhancement by distilling noise-invariant acoustic-semantic representations from noisy inputs to condition an autoregressive decoder-only language model.
-
UniPASE: A Generative Model for Universal Speech Enhancement with High Fidelity and Low Hallucinations
UniPASE extends the low-hallucination PASE framework to universal speech enhancement, restoring seven distortion types at flexible sampling rates with better word-error and speaker-similarity scores than prior generative systems.
-
PhASE-Flow: Phonetic-Conditioned Acoustic Flow Matching in SSL Representation Domain for Speech Enhancement
PhASE-Flow performs phonetic-conditioned acoustic flow matching entirely in SSL representation space for speech enhancement and reports competitive perceptual quality with only four sampling steps.