Let's Fuse Step by Step: A Generative Fusion Decoding Algorithm with LLMs for Robust and Instruction-Aware ASR and OCR
This work addresses robust and instruction-aware ASR and OCR for users needing improved text recognition accuracy, though it is incremental as it builds on existing fusion and LLM techniques.
The paper tackled the problem of integrating large language models into cross-modal text recognition systems for ASR and OCR by proposing a shallow fusion framework that enables seamless fusion across mismatched token spaces, resulting in significant WER reductions of up to 17.7% and surpassing cascaded methods in benchmarks.
We propose "Generative Fusion Decoding" (GFD), a novel shallow fusion framework designed to integrate large language models (LLMs) into cross-modal text recognition systems for automatic speech recognition (ASR) and optical character recognition (OCR). We derive the necessary formulations to enable GFD to operate across mismatched token spaces of different models by calculating likelihood at the byte level, thereby enabling seamless fusion and synchronous progression during the decoding process. GFD is plug-and-play by design, making it readily compatible with various auto-regressive models without the need for any re-training. GFD proves effective for general ASR and OCR tasks through intermediate and frequent interactions with LLMs, surpassing cascaded methods in English and Mandarin benchmarks. In addition, GFD transfers in-context learning abilities of LLMs and allows for adaptive ASR in instruction-aware and long-context settings, yielding significant WER reductions of up to 17.7\%.