{"about":{"non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","site":"https://codewithpapers.app","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page","syntology":{"site":"https://syntology.ai","developers":"https://syntology.ai/developers","mcp":{"server":"https://syntology.ai/mcp","transport":"streamable-http","server_card":"https://syntology.ai/.well-known/mcp/server-card.json","auth":{"type":"trial token, no account","trial_token":"https://syntology.ai/api/oauth/trial/token","method":"POST","docs":"https://syntology.ai/developers"}},"have":"https://syntology.ai/api/graph/have?x=<method, arXiv id or title> (free, answers coverage only)","paper_base":"https://syntology.ai/paper/","atlas_base":"https://app.syntology.ai/?focus="},"machine_readable":[{"url":"https://codewithpapers.app/llms.txt","what":"the machine catalog: every machine-readable file, counted"},{"url":"https://codewithpapers.app/index/manifest.json","what":"paper-to-code index by arXiv id, with Syntology's counts"},{"url":"https://codewithpapers.app/search/manifest.json","what":"site search index (titles, authors) and its files"},{"url":"https://codewithpapers.app/download","what":"bulk files: Syntology's layer, described there"},{"url":"https://codewithpapers.app/build_manifest.json","what":"the build record: inputs, counts, exclusions, probes"}]},"url":"/paper/mambattention-mamba-with-multi-head-attention","title":"MambAttention: Mamba with Multi-Head Attention for Generalizable Single-Channel Speech Enhancement","arxiv_id":"2507.00966","date":"2025-07-01","proceeding":null,"authors":["Nikolai Lund Kühne","Jesper Jensen","Jan Østergaard","Zheng-Hua Tan"],"abstract":"With the advent of new sequence models like Mamba and xLSTM, several studies have shown that these models match or outperform state-of-the-art models in single-channel speech enhancement, automatic speech recognition, and self-supervised audio representation learning. However, prior research has demonstrated that sequence models like LSTM and Mamba tend to overfit to the training set. To address this issue, previous works have shown that adding self-attention to LSTMs substantially improves generalization performance for single-channel speech enhancement. Nevertheless, neither the concept of hybrid Mamba and time-frequency attention models nor their generalization performance have been explored for speech enhancement. In this paper, we propose a novel hybrid architecture, MambAttention, which combines Mamba and shared time- and frequency-multi-head attention modules for generalizable single-channel speech enhancement. To train our model, we introduce VoiceBank+Demand Extended (VB-DemandEx), a dataset inspired by VoiceBank+Demand but with more challenging noise types and lower signal-to-noise ratios. Trained on VB-DemandEx, our proposed MambAttention model significantly outperforms existing state-of-the-art LSTM-, xLSTM-, Mamba-, and Conformer-based systems of similar complexity across all reported metrics on two out-of-domain datasets: DNS 2020 and EARS-WHAM_v2, while matching their performance on the in-domain dataset VB-DemandEx. Ablation studies highlight the role of weight sharing between the time- and frequency-multi-head attention modules for generalization performance. Finally, we explore integrating the shared time- and frequency-multi-head attention modules with LSTM and xLSTM, which yields a notable performance improvement on the out-of-domain datasets. However, our MambAttention model remains superior on both out-of-domain datasets across all reported evaluation metrics.","url_abs":"https://arxiv.org/abs/2507.00966v1","url_pdf":"https://arxiv.org/pdf/2507.00966v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"mambattention-mamba-with-multi-head-attention","repo_url":"https://github.com/NikolaiKyhne/MambAttention","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"mambattention-mamba-with-multi-head-attention","repo_url":"https://github.com/nikolaikyhne/xlstm-senet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"automatic-speech-recognition-2","task_name":"Automatic Speech Recognition"},{"task_slug":"mamba","task_name":"Mamba"},{"task_slug":"speech-enhancement","task_name":"Speech Enhancement"},{"task_slug":"speech-recognition","task_name":"Speech Recognition"},{"task_slug":"speech-recognition-1","task_name":"speech-recognition"}],"methods":[{"method_slug":"lstm","method_name":"LSTM"},{"method_slug":"mamba","method_name":"Mamba"}],"datasets_introduced":[{"slug":"vb-demandex","name":"VB-DemandEx","full_name":""}],"methods_introduced":[],"results":[{"leaderboard":"/sota/speech-enhancement-on-deep-noise-suppression","task":"Speech Enhancement","dataset":"Deep Noise Suppression (DNS) Challenge","model":"MambAttention","rank_in_archive_order":4,"of":36,"metrics":{"ESTOI":"95.9","Number of parameters (M)":"2.33","PESQ-WB":"3.671","SI-SDR-WB":"21.234","SSNR":"15.116"},"uses_additional_data":false},{"leaderboard":"/sota/speech-enhancement-on-deep-noise-suppression","task":"Speech Enhancement","dataset":"Deep Noise Suppression (DNS) Challenge","model":"xLSTM-SENet","rank_in_archive_order":6,"of":36,"metrics":{"ESTOI":"95.4","Number of parameters (M)":"2.20","PESQ-WB":"3.588","SI-SDR-WB":"20.854","SSNR":"14.526"},"uses_additional_data":false},{"leaderboard":"/sota/speech-enhancement-on-vb-demandex","task":"Speech Enhancement","dataset":"VB-DemandEx","model":"MambAttention","rank_in_archive_order":1,"of":4,"metrics":{"ESTOI":"0.801","Number of parameters (M)":"2.33","PESQ (wb)":"3.026","SI-SDR":"16.684","SSNR":"7.674"},"uses_additional_data":false},{"leaderboard":"/sota/speech-enhancement-on-vb-demandex","task":"Speech Enhancement","dataset":"VB-DemandEx","model":"SEMamba","rank_in_archive_order":2,"of":4,"metrics":{"ESTOI":"0.800","Number of parameters (M)":"2.25","PESQ (wb)":"3.002","SI-SDR":"16.593","SSNR":"7.590"},"uses_additional_data":false},{"leaderboard":"/sota/speech-enhancement-on-vb-demandex","task":"Speech Enhancement","dataset":"VB-DemandEx","model":"xLSTM-SENet","rank_in_archive_order":3,"of":4,"metrics":{"ESTOI":"0.795","Number of parameters (M)":"2.20","PESQ (wb)":"2.973","SI-SDR":"16.414","SSNR":"7.933"},"uses_additional_data":false},{"leaderboard":"/sota/speech-enhancement-on-vb-demandex","task":"Speech Enhancement","dataset":"VB-DemandEx","model":"MP-SENet","rank_in_archive_order":4,"of":4,"metrics":{"ESTOI":"0.787","Number of parameters (M)":"2.05","PESQ (wb)":"2.935","SI-SDR":"16.202","SSNR":"7.641"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers","held":null,"reason":"arxiv_id_not_in_atlas","checked_against":{"atlas_ids":549449,"atlas_ids_sha256":"afd72399dafd9906e9062927f70655952385a3d75488558e1f0484e950ab09b7","read_at":"2026-10-05T01:46:21Z"},"by_repo":[{"repo":"NikolaiKyhne/MambAttention","kind":"official","samples":0,"ran":0,"instrument":0,"unverified":0},{"repo":"nikolaikyhne/xlstm-senet","kind":"listed","samples":0,"ran":0,"instrument":0,"unverified":0}],"by_repo_is":"one row per repository in the archive's code links, Syntology's graph links or the samples; samples 0 means no sample is linked to this paper, a statement about Syntology's coverage, not about the repository: with harvested_for_other_papers present, Syntology harvested the repository for that many other papers, and without it, Syntology harvested nothing from it; placed_by_identical_code counts samples whose link records no repository, placed here because identical code was harvested from this repository for another paper; ran counts executions on a synthesized input and instrument counts failures of Syntology's instrument, not of the code"},"code_links_note":null,"arxiv_metadata":null,"syntology_extracted_results":null,"record_sha256":"e2195ebf68b6222a52e8371e73e6f2fd523767375215f4bd68b036b765fdfaec","record_changed_at":"2026-09-28","record_changed_at_basis":"first_hashed"}