Analyzing Error Propagation in Korean Spoken QA with ASR-LLM Cascades
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866913136938844160 |
|---|---|
| author | Jung, Donghyuk Choi, Youngwon |
| author_facet | Jung, Donghyuk Choi, Youngwon |
| contents | We analyze how automatic speech recognition (ASR) errors propagate through ASR-LLM cascades in Korean spoken question answering (SQA), focusing on downstream semantic failures that conventional ASR metrics cannot fully capture. Our analysis shows that the relative downstream degradation caused by ASR errors is consistent across LLMs with different absolute performance, suggesting that cascade degradation largely tracks ASR-stage information loss. We further identify single-character Korean ASR errors as a distinct semantic-failure channel, where the gold answer becomes entirely absent from the downstream prediction despite only a minimal transcription difference. Finally, an auxiliary comparison shows that a large audio language model outperforms an ASR-LLM pipeline with a matched language backbone in noisy Korean SQA, indicating the potential of direct audio input to mitigate transcript-induced information loss. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_17443 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Analyzing Error Propagation in Korean Spoken QA with ASR-LLM Cascades Jung, Donghyuk Choi, Youngwon Computation and Language Sound Audio and Speech Processing We analyze how automatic speech recognition (ASR) errors propagate through ASR-LLM cascades in Korean spoken question answering (SQA), focusing on downstream semantic failures that conventional ASR metrics cannot fully capture. Our analysis shows that the relative downstream degradation caused by ASR errors is consistent across LLMs with different absolute performance, suggesting that cascade degradation largely tracks ASR-stage information loss. We further identify single-character Korean ASR errors as a distinct semantic-failure channel, where the gold answer becomes entirely absent from the downstream prediction despite only a minimal transcription difference. Finally, an auxiliary comparison shows that a large audio language model outperforms an ASR-LLM pipeline with a matched language backbone in noisy Korean SQA, indicating the potential of direct audio input to mitigate transcript-induced information loss. |
| title | Analyzing Error Propagation in Korean Spoken QA with ASR-LLM Cascades |
| topic | Computation and Language Sound Audio and Speech Processing |
| url | https://arxiv.org/abs/2605.17443 |