Panos Georgiou (USC) “Speech Processing & Machine Learning for Behavior Analysis”

When:
October 5, 2018 @ 12:00 pm – 1:15 pm
2018-10-05T12:00:00-04:00
2018-10-05T13:15:00-04:00
Where:
Hackerman Hall B17
3400 N Charles St
Baltimore, MD 21218
USA
Cost:
Free

Abstract

The expression and experience of human behavior are complex and multimodal and characterized by individual and contextual heterogeneity and variability. Speech and spoken language communication cues offer an important means for measuring and modeling human behavior.

Assessment of behavior is also a critical element for behavioral sciences and especially for mental health. Human assessment is often costly, subjective, slow, and biased, especially in atypical cases, which are the ones of greatest interest.

Mental health domain experts can inform development of methods that can assess human communication towards behavior analysis. Expert knowledge can, for example, define and guide behaviors and patterns of interest and can provide exemplars. Machine learning also provides unique opportunities towards data-driven assessment.

In this talk, I am going to present an overview of this convergence towards analysis of couples’ therapy interactions. I will describe specific examples of knowledge and data-driven machine learning methods that can aid mental health experts. I will discuss a range of directions such as outcome prediction and unsupervised learning.

Biography

Panayiotis (Panos) Georgiou is an Assistant Professor in Electrical Engineering and Computer Science at the University of Southern California (USC) and the director of the Signal Processing for Communication Understanding and Behavior Analysis (SCUBA). Panos received his B.A. and M.Eng. degrees (with Honors) from Cambridge University (Pembroke College), Cambridge, U.K.  where he was a Cambridge-Commonwealth Scholar, and his M.Sc. and Ph.D. degrees from the University of Southern California (USC), Los Angeles.

He has worked on and published over 180 papers in the fields of behavioral signal processing, statistical signal processing, alpha stable distributions, speech and multimodal signal processing and interfaces, speech translation, language modeling, immersive sound processing, sound source localization, and speaker identification. He has been a PI and co-PI on federally funded projects notably including DARPA Transtac “SpeechLinks,” NSF “An Integrated Approach to Creating Enriched Speech Translation Systems,” NSF “Quantitative Observational Practice in Family Studies: The case of reactivity,” and DoD “Technologies for assessing suicide risk”.

Papers co-authored with his students have won 3 best paper awards and an interspeech paralinguistics challenge award. He is currently a member of IEEE-SLTC, editor of IEEE Signal Processing Letters, IEEE Signal Processing Magazine, EURASIP Journal on Audio, Speech, and Music Processing, Advances in Artificial Intelligence and served as a guest editor of Computer Speech And Language. He has also served and serves on various conference organizing committees, including technical chair of Interspeech 2016 and general chair of ICMI 2018.

Panos’ work has also been featured in over 100 national and international media outlets such as Washington Post, Telegraph U.K., US News and World report, etc.

His current focus is on behavioral signal processing, multimodal environments, and machine learning for speech applications

Johns Hopkins University

Johns Hopkins University, Whiting School of Engineering

Center for Language and Speech Processing
Hackerman 226
3400 North Charles Street, Baltimore, MD 21218-2608

Center for Language and Speech Processing