2021 IEEE International Conference on Acoustics, Speech and Signal Processing

6-11 June 2021 • Toronto, Ontario, Canada

Extracting Knowledge from Information

2021 IEEE International Conference on Acoustics, Speech and Signal Processing

6-11 June 2021 • Toronto, Ontario, Canada

Extracting Knowledge from Information

Technical Program

Click on the icon to view the manuscript on IEEE XPlore in the IEEE ICASSP 2021 Open Preview.

SPE-21: Speech Recognition 7: Training Methods for End-to-End Modeling

Session Type: Poster
Time: Wednesday, 9 June, 15:30 - 16:15
Location: Gather.Town
Virtual Session: View on Virtual Platform
Session Chair: Karen Livescu, TTI-Chicago
 
 SPE-21.1: TOP-DOWN ATTENTION IN END-TO-END SPOKEN LANGUAGE UNDERSTANDING
         Yixin Chen; University of California, Los Angeles
         Weiyi Lu; Amazon Alexa
         Alejandro Mottini; Amazon Alexa
         Li Erran Li; Amazon Alexa
         Jasha Droppo; Amazon Alexa
         Zheng Du; Amazon Alexa
         Belinda Zeng; Amazon Alexa
 
 SPE-21.2: FINE-TUNING OF PRE-TRAINED END-TO-END SPEECH RECOGNITION WITH GENERATIVE ADVERSARIAL NETWORKS
         Md. Akmal Haidar; Huawei Noah's Ark Lab
         Mehdi Rezagholizadeh; Huawei Noah's Ark Lab
 
 SPE-21.3: A GENERAL MULTI-TASK LEARNING FRAMEWORK TO LEVERAGE TEXT DATA FOR SPEECH TO TEXT TASKS
         Yun Tang; Facebook
         Juan Pino; Facebook
         Changhan Wang; Facebook
         Xutai Ma; Johns Hopkins University
         Dmitriy Genzel; Facebook
 
 SPE-21.4: GAUSSIAN KERNELIZED SELF-ATTENTION FOR LONG SEQUENCE DATA AND ITS APPLICATION TO CTC-BASED SPEECH RECOGNITION
         Yosuke Kashiwagi; Sony Corporation
         Emiru Tsunoo; Sony Corporation
         Shinji Watanabe; Johns Hopkins University
 
 SPE-21.5: LATTICE-FREE MMI ADAPTATION OF SELF-SUPERVISED PRETRAINED ACOUSTIC MODELS
         Apoorv Vyas; Idiap Research Institute and EPFL
         Srikanth Madikeri; Idiap Research Institute
         Hervé Bourlard; Idiap Research Institute
 
 SPE-21.6: INTERMEDIATE LOSS REGULARIZATION FOR CTC-BASED SPEECH RECOGNITION
         Jaesong Lee; Naver Corporation
         Shinji Watanabe; Johns Hopkins University