SPE-42.4
SPEECHMOE2: MIXTURE-OF-EXPERTS MODEL WITH IMPROVED ROUTING
Zhao You, Shulin Feng, Dan Su, Dong Yu, Tencent, China
Session:
Speech Recognition: Acoustic Modeling I
Track:
Speech and Language Processing
Location:
Gather Area C
Presentation Time:
Tue, 10 May, 22:00 - 22:45 China Time (UTC +8)
Tue, 10 May, 14:00 - 14:45 UTC
Tue, 10 May, 14:00 - 14:45 UTC
Session Chair:
Xiadong Cui, IBM
Session SPE-42
SPE-42.1: WORD ORDER DOES NOT MATTER FOR SPEECH RECOGNITION
Vineel Pratap Konduru, Qiantong Xu, Tatiana Likhomanenko, Gabriel Synnaeve, Ronan Collobert, Facebook AI, United States of America
SPE-42.2: CONTRASTIVE SIAMESE NETWORK FOR SEMI-SUPERVISED SPEECH RECOGNITION
Soheil Khorram, Jaeyoung Kim, Anshuman Tripathi, Han Lu, Qian Zhang, Hasim Sak, Google Inc. USA, United States of America
SPE-42.3: SEQUENCE TRANSDUCTION WITH GRAPH-BASED SUPERVISION
Niko Moritz, Facebook AI, United Kingdom of Great Britain and Northern Ireland; Takaaki Hori, Jonathan Le Roux, MERL, United States of America; Shinji Watanabe, Carnegie Mellon University, United States of America
SPE-42.4: SPEECHMOE2: MIXTURE-OF-EXPERTS MODEL WITH IMPROVED ROUTING
Zhao You, Shulin Feng, Dan Su, Dong Yu, Tencent, China
SPE-42.5: SUPERVISED ATTENTION IN SEQUENCE-TO-SEQUENCE MODELS FOR SPEECH RECOGNITION
Gene-Ping Yang, Hao Tang, University of Edinburgh, United Kingdom of Great Britain and Northern Ireland
SPE-42.6: End-to-End Speech Recognition from Federated Acoustic Models
Yan Gao, Pedro P. B. de Gusmao, Nicholas D. Lane, University of Cambridge, United Kingdom of Great Britain and Northern Ireland; Titouan Parcollet, Avignon University, France; Salah Zaiem, Telecom Paris, France; Javier Fernandez-Marques, University of Oxford, United Kingdom of Great Britain and Northern Ireland; Daniel J. Beutel, Adap GmbH; University of Cambridge, Germany