Forgotten Words: Benchmarking NeoBERT for Dementia Detection in Low-Resource Conversational Filipino and English Speech
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866913161762832384 |
|---|---|
| author | Floresca, Rez Samantha Z. Hao, Edric Castel C. Buñales, Hannah Grachiella Temprosa, Chelsea Dominique E. Reyes, Georgianna Z. Chua, Kervin Gabriel L. |
| author_facet | Floresca, Rez Samantha Z. Hao, Edric Castel C. Buñales, Hannah Grachiella Temprosa, Chelsea Dominique E. Reyes, Georgianna Z. Chua, Kervin Gabriel L. |
| contents | Dementia detection from spontaneous speech offers a scalable approach to cognitive screening, yet NLP systems remain predominantly English-centric. This limitation is especially acute in the Philippines, where Filipino-English code-switching is pervasive and no prior work has addressed NLP-based dementia detection. We present the first systematic evaluation of transformer-based dementia detection in Filipino speech and the first assessment of NeoBERT in a clinical NLP setting. To separate language from domain effects, we construct a parallel bilingual dataset of 4,000 DementiaBank-derived transcripts, with Filipino translations produced manually to preserve discourse-level markers of cognitive decline. We evaluate five model families, TF-IDF + LogReg, BERT, NeoBERT, XLM-R, and RoBERTa-Tagalog, under monolingual, zero-shot cross-lingual, and bilingual fine-tuning settings. We find that in-domain performance does not transfer across languages, with English-trained BERT dropping to Macro-F1 = 0.455 on Filipino, and that architectural modernization alone does not improve robustness. Bilingual fine-tuning, however, eliminates cross-lingual degradation across all transformer models, converging to Macro-F1 = 0.969-0.973. These results suggest that multilingual clinical NLP performance is driven primarily by linguistic coverage during training rather than model scale or architecture. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_26007 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Forgotten Words: Benchmarking NeoBERT for Dementia Detection in Low-Resource Conversational Filipino and English Speech Floresca, Rez Samantha Z. Hao, Edric Castel C. Buñales, Hannah Grachiella Temprosa, Chelsea Dominique E. Reyes, Georgianna Z. Chua, Kervin Gabriel L. Computation and Language I.2.7; J.3 Dementia detection from spontaneous speech offers a scalable approach to cognitive screening, yet NLP systems remain predominantly English-centric. This limitation is especially acute in the Philippines, where Filipino-English code-switching is pervasive and no prior work has addressed NLP-based dementia detection. We present the first systematic evaluation of transformer-based dementia detection in Filipino speech and the first assessment of NeoBERT in a clinical NLP setting. To separate language from domain effects, we construct a parallel bilingual dataset of 4,000 DementiaBank-derived transcripts, with Filipino translations produced manually to preserve discourse-level markers of cognitive decline. We evaluate five model families, TF-IDF + LogReg, BERT, NeoBERT, XLM-R, and RoBERTa-Tagalog, under monolingual, zero-shot cross-lingual, and bilingual fine-tuning settings. We find that in-domain performance does not transfer across languages, with English-trained BERT dropping to Macro-F1 = 0.455 on Filipino, and that architectural modernization alone does not improve robustness. Bilingual fine-tuning, however, eliminates cross-lingual degradation across all transformer models, converging to Macro-F1 = 0.969-0.973. These results suggest that multilingual clinical NLP performance is driven primarily by linguistic coverage during training rather than model scale or architecture. |
| title | Forgotten Words: Benchmarking NeoBERT for Dementia Detection in Low-Resource Conversational Filipino and English Speech |
| topic | Computation and Language I.2.7; J.3 |
| url | https://arxiv.org/abs/2605.26007 |