[{"citation":{"short":"K. Kinoshita, T.C. von Neumann, M. Delcroix, T. Nakatani, R. Haeb-Umbach, in: Proc. Interspeech 2020, 2020, pp. 2652–2656.","ama":"Kinoshita K, von Neumann TC, Delcroix M, Nakatani T, Haeb-Umbach R. Multi-Path RNN for Hierarchical Modeling of Long Sequential Data and its Application to Speaker Stream Separation. In: <i>Proc. Interspeech 2020</i>. ; 2020:2652-2656. doi:<a href=\"https://doi.org/10.21437/Interspeech.2020-2388\">10.21437/Interspeech.2020-2388</a>","mla":"Kinoshita, Keisuke, et al. “Multi-Path RNN for Hierarchical Modeling of Long Sequential Data and Its Application to Speaker Stream Separation.” <i>Proc. Interspeech 2020</i>, 2020, pp. 2652–56, doi:<a href=\"https://doi.org/10.21437/Interspeech.2020-2388\">10.21437/Interspeech.2020-2388</a>.","ieee":"K. Kinoshita, T. C. von Neumann, M. Delcroix, T. Nakatani, and R. Haeb-Umbach, “Multi-Path RNN for Hierarchical Modeling of Long Sequential Data and its Application to Speaker Stream Separation,” in <i>Proc. Interspeech 2020</i>, 2020, pp. 2652–2656.","apa":"Kinoshita, K., von Neumann, T. C., Delcroix, M., Nakatani, T., &#38; Haeb-Umbach, R. (2020). Multi-Path RNN for Hierarchical Modeling of Long Sequential Data and its Application to Speaker Stream Separation. In <i>Proc. Interspeech 2020</i> (pp. 2652–2656). <a href=\"https://doi.org/10.21437/Interspeech.2020-2388\">https://doi.org/10.21437/Interspeech.2020-2388</a>","bibtex":"@inproceedings{Kinoshita_von Neumann_Delcroix_Nakatani_Haeb-Umbach_2020, title={Multi-Path RNN for Hierarchical Modeling of Long Sequential Data and its Application to Speaker Stream Separation}, DOI={<a href=\"https://doi.org/10.21437/Interspeech.2020-2388\">10.21437/Interspeech.2020-2388</a>}, booktitle={Proc. Interspeech 2020}, author={Kinoshita, Keisuke and von Neumann, Thilo Christoph and Delcroix, Marc and Nakatani, Tomohiro and Haeb-Umbach, Reinhold}, year={2020}, pages={2652–2656} }","chicago":"Kinoshita, Keisuke, Thilo Christoph von Neumann, Marc Delcroix, Tomohiro Nakatani, and Reinhold Haeb-Umbach. “Multi-Path RNN for Hierarchical Modeling of Long Sequential Data and Its Application to Speaker Stream Separation.” In <i>Proc. Interspeech 2020</i>, 2652–56, 2020. <a href=\"https://doi.org/10.21437/Interspeech.2020-2388\">https://doi.org/10.21437/Interspeech.2020-2388</a>."},"user_id":"59789","type":"conference","status":"public","publication":"Proc. Interspeech 2020","oa":1,"accept":"1","page":"2652-2656","year":"2020","date_created":"2020-12-16T14:15:24Z","abstract":[{"text":"Recently, the source separation performance was greatly improved by time-domain audio source separation based on dual-path recurrent neural network (DPRNN). DPRNN is a simple but effective model for a long sequential data. While DPRNN is quite efficient in modeling a sequential data of the length of an utterance, i.e., about 5 to 10 second data, it is harder to apply it to longer sequences such as whole conversations consisting of multiple utterances. It is simply because, in such a case, the number of time steps consumed by its internal module called inter-chunk RNN becomes extremely large. To mitigate this problem, this paper proposes a multi-path RNN (MPRNN), a generalized version of DPRNN, that models the input data in a hierarchical manner. In the MPRNN framework, the input data is represented at several (>_ 3) time-resolutions, each of which is modeled by a specific RNN sub-module. For example, the RNN sub-module that deals with the finest resolution may model temporal relationship only within a phoneme, while the RNN sub-module handling the most coarse resolution may capture only the relationship between utterances such as speaker information. We perform experiments using simulated dialogue-like mixtures and show that MPRNN has greater model capacity, and it outperforms the current state-of-the-art DPRNN framework especially in online processing scenarios.","lang":"eng"}],"_id":"20766","cc_license":"' http://creativecommons.org/publicdomain/zero/1.0/'","author":[{"first_name":"Keisuke","last_name":"Kinoshita","full_name":"Kinoshita, Keisuke"},{"full_name":"von Neumann, Thilo Christoph","id":"49870","first_name":"Thilo Christoph","last_name":"von Neumann"},{"first_name":"Marc","last_name":"Delcroix","full_name":"Delcroix, Marc"},{"last_name":"Nakatani","first_name":"Tomohiro","full_name":"Nakatani, Tomohiro"},{"id":"242","last_name":"Haeb-Umbach","first_name":"Reinhold","full_name":"Haeb-Umbach, Reinhold"}],"ddc":["000"],"file":[{"creator":"huesera","relation":"main_file","content_type":"application/pdf","file_id":"20767","request_a_copy":0,"access_level":"open_access","file_size":1725219,"date_created":"2020-12-16T14:16:32Z","file_name":"INTERSPEECH_2020_vonNeumann1_Paper.pdf","date_updated":"2020-12-16T14:16:32Z"}],"doi":"10.21437/Interspeech.2020-2388","file_date_updated":"2020-12-16T14:16:32Z","department":[{"_id":"54"}],"language":[{"iso":"eng"}],"project":[{"_id":"52","name":"Computing Resources Provided by the Paderborn Center for Parallel Computing"}],"date_updated":"2020-12-16T14:17:23Z","title":"Multi-Path RNN for Hierarchical Modeling of Long Sequential Data and its Application to Speaker Stream Separation"}]