Can genera- tive large language models perform ASR error correction?

representative citing papers

citing papers explorer

Showing 5 of 5 citing papers.

• Can Large Language Models Reliably Correct Errors in Low-Resource ASR? A Contamination-Aware Case Study on West Frisian cs.CL · 2026-05-19 · conditional · none · ref 11

  LLM generative error correction improves low-resource Frisian ASR performance, with comparable gains on a contamination-controlled offline dataset confirming true correction ability.

• Phonemes vs. Projectors: An Investigation of Speech-Language Interfaces for LLM-based ASR eess.AS · 2026-04-10 · unverdicted · none · ref 16

  Phoneme-based interfaces match or surpass projector-based ones for LLM ASR, especially in low-resource languages, and a BPE-phoneme hybrid offers additional improvements.

• Reducing Prompt Sensitivity in LLM-based Speech Recognition Through Learnable Projection eess.AS · 2026-01-28 · conditional · none · ref 12

  A learnable prompt projector added to LLM-based ASR reduces prompt sensitivity, lowers performance variability, and beats the best fixed prompts on four datasets.

• Read What You Hear: Reference-Free Hypotheses Evaluation with Acoustic Discrepancy eess.AS · 2026-06-03 · unverdicted · none · ref 15

  READ is a reference-free ASR hypothesis scorer that measures acoustic discrepancy via conditional likelihood from a pretrained auto-regressive TTS model and yields up to 20% relative error rate reduction when used for refinement.

• Error-Aware TF-IDF Retrieval-Augmented Generation for ASR Error Correction cs.CL · 2026-06-19 · unverdicted · none · ref 7

  Introduces error-aware TF-IDF RAG that raises error-aware hit rate from 53.7% to 90.9% and lowers WER from 23.06% to 18.83% on Persian FLEURS data.