{"type":"ScholarlyArticle","title":"M3ANet: Multi-scale and Multi-Modal Alignment Network for Brain-Assisted Target Speaker Extraction","authors":[{"name":"Cunhang Fan","slug":"cunhang-fan","given":"Cunhang","family":"Fan","affiliation":""},{"name":"Ying Chen","slug":"ying-chen","given":"Ying","family":"Chen","affiliation":""},{"name":"Jian Zhou","slug":"jian-zhou","given":"Jian","family":"Zhou","affiliation":""},{"name":"Zexu Pan","slug":"zexu-pan","given":"Zexu","family":"Pan","affiliation":""},{"name":"Jingjing Zhang","slug":"jingjing-zhang","given":"Jingjing","family":"Zhang","affiliation":""},{"name":"Youdian Gao","slug":"youdian-gao","given":"Youdian","family":"Gao","affiliation":""},{"name":"Xiaoke Yang","slug":"xiaoke-yang","given":"Xiaoke","family":"Yang","affiliation":""},{"name":"Zhengqi Wen","slug":"zhengqi-wen","given":"Zhengqi","family":"Wen","affiliation":""},{"name":"Zhao Lv","slug":"zhao-lv","given":"Zhao","family":"Lv","affiliation":""}],"abstract":"The brain-assisted target speaker extraction (TSE) aims to extract the attended speech from mixed speech by utilizing the brain neural activities, for example Electroencephalography (EEG). However, existing models overlook the issue of temporal misalignment between speech and EEG modalities, which hampers TSE performance. In addition, the speech encoder in current models typically uses basic temporal operations (e.g., one-dimensional convolution), which are unable to effectively extract target speaker information. To address these issues, this paper proposes a multi-scale and multi-modal alignment network (M3ANet) for brain-assisted TSE. Specifically, to eliminate the temporal inconsistency between EEG and speech modalities, the modal alignment module that uses a contrastive learning strategy is applied to align the temporal features of both modalities. Additionally, to fully extract speech information, multi-scale convolutions with GroupMamba modules are used as the speech encoder, which scans speech features at each scale from different directions, enabling the model to capture deep sequence information. Experimental results on three publicly available datasets show that the proposed model outperforms current state-of-the-art methods across various evaluation metrics, highlighting the effectiveness of our proposed method. The source code is available at: https://github.com/fchest/M3ANet.","doi":"10.24963/ijcai.2025/894","url":"https://ijcai.org/proceedings/2025/894","pdf":"https://ijcai.org/proceedings/2025/0894.pdf","markdown":"https://ijcai.org/proceedings/2025/894.md","bibtex":"https://ijcai.org/proceedings/2025/bibtex/894","volume":{"slug":"2025","title":"Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2025","url":"https://ijcai.org/proceedings/2025/","year":2025,"location":"Montreal, Canada","dates":"16-22 August 2025","editors":"James Kwok","isbn":"978-1-956792-06-5","publisher":"International Joint Conferences on Artificial Intelligence Organization","doi_prefix":"10.24963/ijcai.2025"},"section":"Natural Language Processing","track":"Main Track","number":894,"pages":{"first":8040,"last":8048},"keywords":["Natural Language Processing: NLP: Speech","Humans and AI: HAI: Human-computer interaction"],"videos":[],"references":[],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}