Context-Aware Dynamic Chunking for Streaming Tibetan Speech Recognition
This work addresses the challenge of accurate and low-latency speech recognition for Tibetan speakers, which is an incremental improvement in a domain-specific context.
The paper tackled the problem of streaming speech recognition for Amdo Tibetan by proposing a context-aware dynamic chunking mechanism, achieving a word error rate of 6.23% with a 48.15% relative improvement over the baseline.
In this work, we propose a streaming speech recognition framework for Amdo Tibetan, built upon a hybrid CTC/Atten-tion architecture with a context-aware dynamic chunking mechanism. The proposed strategy adaptively adjusts chunk widths based on encoding states, enabling flexible receptive fields, cross-chunk information exchange, and robust adaptation to varying speaking rates, thereby alleviating the context truncation problem of fixed-chunk methods. To further capture the linguistic characteristics of Tibetan, we construct a lexicon grounded in its orthographic principles, providing linguistically motivated modeling units. During decoding, an external language model is integrated to enhance semantic consistency and improve recognition of long sentences. Experimental results show that the proposed framework achieves a word error rate (WER) of 6.23% on the test set, yielding a 48.15% relative improvement over the fixed-chunk baseline, while significantly reducing recognition latency and maintaining performance close to global decoding.