Hindi

The 'Hindi' dataset/benchmark contains transcriptions in Hindi and is used to evaluate the performance of Automatic Speech Recognition (ASR) systems through fine-grained Part-of-Speech (PoS)-wise error characterization and alignment analysis.

Papers using Hindi (12)

Breaking the Script Barrier: Enabling Automatic Alignment for PoS-based ASR Error Analysis in Non-Latin Scripts2026Factors affecting ASR performance: A study using state of the art ASR models in Indic Languages2026An Analysis of the Effectiveness of Synthetic Speech Data for ASR Fine-tuning in Selected Indic Languages2026Mind the Pause: Disfluency-Aware Objective Tuning for Multilingual Speech Correction with LLMs2026PSP: An Interpretable Per-Dimension Accent Benchmark for Indic Text-to-Speech2026Dynamic Multi-Expert Projectors with Stabilized Routing for Multilingual Speech Recognition2026Unsupervised Domain Adaptation Schemes for Building ASR in Low-resource Languages2021 · 5 citesData and knowledge-driven approaches for multilingual training to improve the performance of speech recognition systems of Indian languages2022 · 3 citesAdversarial Training For Low-Resource Disfluency Correction2023 · 3 citesPredicting positive transfer for improved low-resource speech recognition using acoustic pseudo-tokens2024 · 2 citesVECL-TTS: Voice identity and Emotional style controllable Cross-Lingual Text-to-Speech2024Advancing Topic Segmentation of Broadcasted Speech with Multilingual Semantic Embeddings2024