Pith. sign in

Rethinking evaluation in ASR: Are our models robust enough? InProc

Tatiana Likhomanenko, Qiantong Xu, Vineel Pratap, Paden Tomasello, Jacob Kahn, Gilad Avidov, Ronan Collobert, Gabriel Synnaeve · 2021 · DOI 10.21437/interspeech.2021-1758

1 Pith paper cite this work, alongside 12 external citations. Polarity classification is still indexing.

1 Pith paper citing it
12 external citations · OpenAlex
open at publisher browse 1 citing papers

fields

cs.CL 1

years

2026 1

verdicts

CONDITIONAL 1

representative citing papers

Voice Memory for Agentic Speech Recognition

cs.CL · 2026-07-29 · conditional · novelty 6.0

Score-gated text memories let a frozen LLM corrector decide when not to edit ASR hypotheses, cutting weighted WER from 8.36% to 7.52% without weight updates.

citing papers explorer

Showing 1 of 1 citing paper.

  • Voice Memory for Agentic Speech Recognition cs.CL · 2026-07-29 · conditional · none · ref 33

    Score-gated text memories let a frozen LLM corrector decide when not to edit ASR hypotheses, cutting weighted WER from 8.36% to 7.52% without weight updates.