The reviewed record of science sign in
Pith

arxiv: 2507.12701 · v1 · pith:LYRY7COK · submitted 2025-07-17 · cs.SD · cs.AI· eess.AS

Task-Specific Audio Coding for Machines: Machine-Learned Latent Features Are Codes for That Machine

Reviewed by Pithpith:LYRY7COKopen to challenge →

classification cs.SD cs.AIeess.AS
keywords audiodownstreammodelspeechacombitratescodingefficient
0
0 comments X
read the original abstract

Neural audio codecs, leveraging quantization algorithms, have significantly impacted various speech/audio tasks. While high-fidelity reconstruction is paramount for human perception, audio coding for machines (ACoM) prioritizes efficient compression and downstream task performance, disregarding perceptual nuances. This work introduces an efficient ACoM method that can compress and quantize any chosen intermediate feature representation of an already trained speech/audio downstream model. Our approach employs task-specific loss guidance alongside residual vector quantization (RVQ) losses, providing ultra-low bitrates (i.e., less than 200 bps) with a minimal loss of the downstream model performance. The resulting tokenizer is adaptable to various bitrates and model sizes for flexible deployment. Evaluated on automatic speech recognition and audio classification, our method demonstrates its efficacy and potential for broader task and architectural applicability through appropriate regularization.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.