In-Context Learning Boosts Speech Recognition via Human-like Adaptation to Speakers and Language Varieties
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913850689847296 |
|---|---|
| author | Roll, Nathan Graham, Calbert Tatsumi, Yuka Nguyen, Kim Tien Sumner, Meghan Jurafsky, Dan |
| author_facet | Roll, Nathan Graham, Calbert Tatsumi, Yuka Nguyen, Kim Tien Sumner, Meghan Jurafsky, Dan |
| contents | Human listeners readily adjust to unfamiliar speakers and language varieties through exposure, but do these adaptation benefits extend to state-of-the-art spoken language models? We introduce a scalable framework that allows for in-context learning (ICL) in Phi-4 Multimodal using interleaved task prompts and audio-text pairs, and find that as few as 12 example utterances (~50 seconds) at inference time reduce word error rates by a relative 19.7% (1.2 pp.) on average across diverse English corpora. These improvements are most pronounced in low-resource varieties, when the context and target speaker match, and when more examples are provided--though scaling our procedure yields diminishing marginal returns to context length. Overall, we find that our novel ICL adaptation scheme (1) reveals a similar performance profile to human listeners, and (2) demonstrates consistent improvements to automatic speech recognition (ASR) robustness across diverse speakers and language backgrounds. While adaptation succeeds broadly, significant gaps remain for certain varieties, revealing where current models still fall short of human flexibility. We release our prompts and code on GitHub. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_14887 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | In-Context Learning Boosts Speech Recognition via Human-like Adaptation to Speakers and Language Varieties Roll, Nathan Graham, Calbert Tatsumi, Yuka Nguyen, Kim Tien Sumner, Meghan Jurafsky, Dan Computation and Language Sound Audio and Speech Processing Human listeners readily adjust to unfamiliar speakers and language varieties through exposure, but do these adaptation benefits extend to state-of-the-art spoken language models? We introduce a scalable framework that allows for in-context learning (ICL) in Phi-4 Multimodal using interleaved task prompts and audio-text pairs, and find that as few as 12 example utterances (~50 seconds) at inference time reduce word error rates by a relative 19.7% (1.2 pp.) on average across diverse English corpora. These improvements are most pronounced in low-resource varieties, when the context and target speaker match, and when more examples are provided--though scaling our procedure yields diminishing marginal returns to context length. Overall, we find that our novel ICL adaptation scheme (1) reveals a similar performance profile to human listeners, and (2) demonstrates consistent improvements to automatic speech recognition (ASR) robustness across diverse speakers and language backgrounds. While adaptation succeeds broadly, significant gaps remain for certain varieties, revealing where current models still fall short of human flexibility. We release our prompts and code on GitHub. |
| title | In-Context Learning Boosts Speech Recognition via Human-like Adaptation to Speakers and Language Varieties |
| topic | Computation and Language Sound Audio and Speech Processing |
| url | https://arxiv.org/abs/2505.14887 |