Early-Signal Detector

model · first seen 7 Oct 2026

speech language model

Also: SpeechLM, SpeechLLM, generative spoken language modeling

A language model that operates directly on raw speech audio rather than transcribed text.

Mentions per week

0519 May11 Aug10 NovPapersCodeDiscussion

PapersCodeDiscussion

Earliest sources

Latest sources

Mentioned together with

synthetic data augmentationimportance samplingsurprisal theoryspeech separationneural audio codecstreaming speech recognition and synthesislow-resource language modelingLLaMA-Omnitool callingknowledge distillation