Search papers, labs, and topics across Lattice.
To address privacy constraints that prevent audio-based error correction in production speech pipelines, the authors curate DasanCallDial鈥攁 115k-utterance real-world Korean call center benchmark鈥攁nd evaluate text-only ASR post-editing. They construct Detector-Gated Contextual Span Correction (DCSC), a framework that couples an encoder-based token-level error detector with a language model trained for fine-grained span correction using dialogue history. This multi-granularity pipeline establishes state-of-the-art performance on Korean conversational transcripts, markedly outperforming prompting-based general large language models on error-sparse text.
When privacy rules forbid raw audio access, general LLMs falter on sparse ASR transcription mistakes鈥攎aking two-stage, detector-gated span correction the far more effective paradigm.
Automatic Speech Recognition (ASR) technology is fundamental to customer service automation and large-scale transcription. However, even advanced ASR models exhibit inevitable errors in complex real-world environments such as call center conversations. When privacy restrictions preclude audio access, error correction must rely on text-based post-editing. Existing text-only approaches face significant challenges in low-resource languages, mainly due to a critical scarcity of annotated corpora and tailored correction methodologies. For Korean, this resource gap is particularly pronounced, as existing resources are predominantly designed for ASR training rather than text-based error correction. To address this, we introduce DasanCallDial, the first large-scale Korean benchmark dataset specifically curated for dialogue-level ASR error correction. Derived from genuine call center interactions, it comprises 1,974 dialogues with 115,460 utterances. Leveraging this resource, we propose Detector-Gated Contextual Span Correction (DCSC), a text-only post-editing framework for error-sparse Korean speech recognition transcripts. DCSC combines an encoder-based detector that first performs token-level error detection, followed by a language model-based corrector trained to rectify fine-grained span-level errors. Additionally, we employ dialogue-level context augmentation to enable the model to leverage discourse history for disambiguation. By employing multi-level granularity, our method achieves state-of-the-art performance, effectively overcoming the limitations of general LLMs in low-resource settings.
