ASME Westchester County Section: Emotion Detection using Transfer Learning from Speech and Speaker Recognition Deep Neural Net Model

Description

Joint Meeting: The ASME Westchester County Section, SME Chapter 216, and IEEE Tappan Zee Subsection presents Emotion Detection using Transfer Learning from Speech and Speaker Recognition Deep Neural Net Model, presented by Dr. Homayoon Beigi, President of Recognition Technologies, Inc. and Adjunct Prof. of Computer Science and Mechanical Engineering, Columbia University. Automated emotion recognition is a challenging task: emotion is abstract, its expression varies across different modalities, the affect of words is context-dependent, and it lacks large datasets. To address these issues, an emotion detection system was built by transfer-learning from speech recognition, speaker recognition, and textual analysis. Join us to learn more about this topic!