Answers are grounded in this paper and cite their sources.
Introduces a post-ASR correction model for Rajasthani, improving accuracy in low-resource settings, implying enhanced language processing capabilities.
State-of-the-art multilingual Automatic Speech Recognition (ASR) models produce systematic errors when applied to low-resource languages like Rajasthani, for which they lack dedicated training data. This paper addresses this challenge by introducing a post-ASR correction framework that leverages the complementary error patterns in the outputs (termed as views) from two distinct models: Whisper-large-v3 and MMS-1B-All. We propose a multi-view 1 , character-level sequence-to-sequence (Seq2Seq) model that uses a gated fusion mechanism to dynamically weigh information from the two ASR outputs. On a new benchmark created from the IndicTTS Rajasthani corpus, our gated model achieves a Character Error Rate (CER) of 7.86% and a Word Error Rate (WER) of 29.98%. This outperforms the best single-view baselines (8.01% CER and 30.33% WER), simple multi-view concatenation (8.21% CER and 30.05% WER), as well as Llama-3.2-3B and mBART-50-large, both fine-tuned on Whisper and MMS inputs. It also surpasses powerful Large Language Models (LLMs) like GPT-4o and Gemini 2.5 Pro in a zero-shot setting. This work establishes the first baseline for post-ASR correction in Rajasthani, demonstrating that a compact, specialized model is more effective than general-purpose LLMs for this targeted, low-resource task.
No takes yet. Share an insight, caveat, or question.
Bhandari et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: