Salvato in:
| Autori principali: | , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2506.20243 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866916811461623808 |
|---|---|
| author | Wade, Papa Séga Andries, Mihai Kanellos, Ioannis Moudenc, Thierry |
| author_facet | Wade, Papa Séga Andries, Mihai Kanellos, Ioannis Moudenc, Thierry |
| contents | Automatic fluency assessment (AFA) remains challenging, particularly in capturing speech rhythm, pauses, and disfluencies in non-native speakers. We introduce a chunk-based approach integrating self-supervised learning (SSL) models (Wav2Vec2, HuBERT, and WavLM) selected for their complementary strengths in phonetic, prosodic, and noisy speech modeling, with a hierarchical CNN-BiLSTM framework. Speech is segmented into breath-group chunks using Silero voice activity detection (Silero-VAD), enabling fine-grained temporal analysis while mitigating over-segmentation artifacts. SSL embeddings are fused via a learnable weighted mechanism, balancing acoustic and linguistic features, and enriched with chunk-level fluency markers (e.g., speech rate, pause durations, n-gram repetitions). The CNN-BiLSTM captures local and long-term dependencies across chunks. Evaluated on Avalinguo and Speechocean762, our approach improves F1-score by 2.8 and Pearson correlation by 6.2 points over single SSL baselines on Speechocean762, with gains of 4.2 F1-score and 4.0 Pearson points on Avalinguo, surpassing Pyannote.audio-based segmentation baselines. These findings highlight chunk-based multi-SSL fusion for robust fluency evaluation, though future work should explore generalization to dialects with irregular prosody. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_20243 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | CBF-AFA: Chunk-Based Multi-SSL Fusion for Automatic Fluency Assessment Wade, Papa Séga Andries, Mihai Kanellos, Ioannis Moudenc, Thierry Computation and Language Artificial Intelligence Audio and Speech Processing Automatic fluency assessment (AFA) remains challenging, particularly in capturing speech rhythm, pauses, and disfluencies in non-native speakers. We introduce a chunk-based approach integrating self-supervised learning (SSL) models (Wav2Vec2, HuBERT, and WavLM) selected for their complementary strengths in phonetic, prosodic, and noisy speech modeling, with a hierarchical CNN-BiLSTM framework. Speech is segmented into breath-group chunks using Silero voice activity detection (Silero-VAD), enabling fine-grained temporal analysis while mitigating over-segmentation artifacts. SSL embeddings are fused via a learnable weighted mechanism, balancing acoustic and linguistic features, and enriched with chunk-level fluency markers (e.g., speech rate, pause durations, n-gram repetitions). The CNN-BiLSTM captures local and long-term dependencies across chunks. Evaluated on Avalinguo and Speechocean762, our approach improves F1-score by 2.8 and Pearson correlation by 6.2 points over single SSL baselines on Speechocean762, with gains of 4.2 F1-score and 4.0 Pearson points on Avalinguo, surpassing Pyannote.audio-based segmentation baselines. These findings highlight chunk-based multi-SSL fusion for robust fluency evaluation, though future work should explore generalization to dialects with irregular prosody. |
| title | CBF-AFA: Chunk-Based Multi-SSL Fusion for Automatic Fluency Assessment |
| topic | Computation and Language Artificial Intelligence Audio and Speech Processing |
| url | https://arxiv.org/abs/2506.20243 |