Index-ASR Technical Report

TL;DR

Index-ASR is a large-scale LLM-based speech recognition system that improves robustness against hallucinations and enables customizable hotword recognition by integrating contextual and background noise data.

Contribution

The paper introduces Index-ASR, a novel LLM-based ASR system that enhances robustness and supports flexible contextual customization, addressing key limitations of existing models.

Findings

Achieves strong performance on open-source benchmarks.

Demonstrates robustness in real-world scenarios.

Supports customizable hotword recognition.

Abstract

Automatic speech recognition (ASR) has witnessed remarkable progress in recent years, largely driven by the emergence of LLM-based ASR paradigm. Despite their strong performance on a variety of open-source benchmarks, existing LLM-based ASR systems still suffer from two critical limitations. First, they are prone to hallucination errors, often generating excessively long and repetitive outputs that are not well grounded in the acoustic input. Second, they provide limited support for flexible and fine-grained contextual customization. To address these challenges, we propose Index-ASR, a large-scale LLM-based ASR system designed to simultaneously enhance robustness and support customizable hotword recognition. The core idea of Index-ASR lies in the integration of LLM and large-scale training data enriched with background noise and contextual information. Experimental results show that our Index-ASR achieves strong performance on both open-source benchmarks and in-house test sets, highlighting its robustness and practicality for real-world ASR applications.