{"ID":2896922,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2507.05729","arxiv_id":"2507.05729","title":"Non-Intrusive Binaural Speech Intelligibility Prediction Using Mamba for Hearing-Impaired Listeners","abstract":"Speech intelligibility prediction (SIP) models have been used as objective metrics to assess intelligibility for hearing-impaired (HI) listeners. In the Clarity Prediction Challenge 2 (CPC2), non-intrusive binaural SIP models based on transformers showed high prediction accuracy. However, the self-attention mechanism theoretically incurs high computational and memory costs, making it a bottleneck for low-latency, power-efficient devices. This may also degrade the temporal processing of binaural SIPs. Therefore, we propose Mamba-based SIP models instead of transformers for the temporal processing blocks. Experimental results show that our proposed SIP model achieves competitive performance compared to the baseline while maintaining a relatively small number of parameters. Our analysis suggests that the SIP model based on bidirectional Mamba effectively captures contextual and spatial speech information from binaural signals.","short_abstract":"Speech intelligibility prediction (SIP) models have been used as objective metrics to assess intelligibility for hearing-impaired (HI) listeners. In the Clarity Prediction Challenge 2 (CPC2), non-intrusive binaural SIP models based on transformers showed high prediction accuracy. However, the self-attention mechanism t...","url_abs":"https://arxiv.org/abs/2507.05729","url_pdf":"https://arxiv.org/pdf/2507.05729v1","authors":"[\"Katsuhiko Yamamoto\",\"Koichi Miyazaki\"]","published":"2025-07-08T07:23:03Z","proceeding":"cs.SD","tasks":"[\"cs.SD\",\"eess.AS\"]","methods":"[\"Transformer\"]","has_code":false}