AIBearisharXiv โ CS AI ยท 15h ago6/10
๐ง
The Cascade Equivalence Hypothesis: When Do Speech LLMs Behave Like ASR$\rightarrow$LLM Pipelines?
Research reveals that speech LLMs don't perform significantly better than traditional ASRโLLM pipelines in most deployed scenarios. The study shows speech LLMs essentially function as expensive cascades that perform worse under noisy conditions, with advantages reversing by up to 7.6% at 0dB noise levels.
$LLM