DeCRED: Decoder-Centric Regularization for Encoder-Decoder Based Speech Recognition
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909734364250112 |
|---|---|
| author | Polok, Alexander Kesiraju, Santosh Beneš, Karel Yusuf, Bolaji Burget, Lukáš Černocký, Jan |
| author_facet | Polok, Alexander Kesiraju, Santosh Beneš, Karel Yusuf, Bolaji Burget, Lukáš Černocký, Jan |
| contents | This paper presents a simple yet effective regularization for the internal language model induced by the decoder in encoder-decoder ASR models, thereby improving robustness and generalization in both in- and out-of-domain settings. The proposed method, Decoder-Centric Regularization in Encoder-Decoder (DeCRED), adds auxiliary classifiers to the decoder, enabling next token prediction via intermediate logits. Empirically, DeCRED reduces the mean internal LM BPE perplexity by 36.6% relative to 11 test sets. Furthermore, this translates into actual WER improvements over the baseline in 5 of 7 in-domain and 3 of 4 out-of-domain test sets, reducing macro WER from 6.4% to 6.3% and 18.2% to 16.2%, respectively. On TEDLIUM3, DeCRED achieves 7.0% WER, surpassing the baseline and encoder-centric InterCTC regularization by 0.6% and 0.5%, respectively. Finally, we compare DeCRED with OWSM v3.1 and Whisper-medium, showing competitive WERs despite training on much less data with fewer parameters. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_08938 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | DeCRED: Decoder-Centric Regularization for Encoder-Decoder Based Speech Recognition Polok, Alexander Kesiraju, Santosh Beneš, Karel Yusuf, Bolaji Burget, Lukáš Černocký, Jan Audio and Speech Processing This paper presents a simple yet effective regularization for the internal language model induced by the decoder in encoder-decoder ASR models, thereby improving robustness and generalization in both in- and out-of-domain settings. The proposed method, Decoder-Centric Regularization in Encoder-Decoder (DeCRED), adds auxiliary classifiers to the decoder, enabling next token prediction via intermediate logits. Empirically, DeCRED reduces the mean internal LM BPE perplexity by 36.6% relative to 11 test sets. Furthermore, this translates into actual WER improvements over the baseline in 5 of 7 in-domain and 3 of 4 out-of-domain test sets, reducing macro WER from 6.4% to 6.3% and 18.2% to 16.2%, respectively. On TEDLIUM3, DeCRED achieves 7.0% WER, surpassing the baseline and encoder-centric InterCTC regularization by 0.6% and 0.5%, respectively. Finally, we compare DeCRED with OWSM v3.1 and Whisper-medium, showing competitive WERs despite training on much less data with fewer parameters. |
| title | DeCRED: Decoder-Centric Regularization for Encoder-Decoder Based Speech Recognition |
| topic | Audio and Speech Processing |
| url | https://arxiv.org/abs/2508.08938 |