NU SEDS Research seminar “Intention Estimation: A Live Feeling Approach” by Professor Martin Lukac
Nazarbayev University’s School of Engineering and Digital Sciences and National Laboratory Astana are delighted to invite you to the Research seminar “Intention Estimation: A Live Feeling Approach” by Professor Martin Lukac, Associate Professor, Department of Computer Science, NU SEDS.
Date: 29 March 2023, Wednesday
Time: 2 pm
Location: Block 3, Room 3.416
Registration form for external participants
Martin Lukac obtained his B.Sc. Degree in Biology from Universite Bordeaux I,, his M.Sc. Degree in Cognitive Sciences from Ecole Polytechnique and University Paris Jussieux, France, and his Master’s and Ph.D. Degree in Computer Engineering from Portland State University, USA in 2009. From 2009 till 2014 he was Assistant Professor at Tohoku University, Sendai, Japan where his main area of research was intelligent robotics and novel-concept algorithm development. From 2015 till 2016 he had a position of Assistant Professor and since 2016 he is an associate professor in the department of Computer Science at Nazarbayev University, Kazakhstan. His interests are quantum computing, artificial intelligence, causal reasoning and neural networks optimization. His main interests range from Artificial Intelligence, through emotional computing and reasoning, machine learning to quantum computing and design of hardware for Artificial Intelligence. He pioneered evolutionary quantum logic synthesis and currently he is working in the area of adaptive robotic vision, meta-learning, Human-Computer Interaction and Deep Neural Network Optimization with the target being real world problems. He has published more than 80 Journal and Conference articles obtaining along the way two best paper awards. Currently he is working on several project related to neural cryptography, real-time human and sports prediction, high level neural network optimizations and the design of quantum algorithms.
ABSTRACT: Efficient Human-Computer Interaction requires the computers or the intelligent agent to understand verbal as well as non-verbal human communication. Up to 60% human-to-human communication is nonverbal and symbolic. In order to understand this requirement we study the problem of human intention estimation in the context of personalized content delivery. For this we propose a construction of the Live-Feeling Communication framework, where an intelligent agent delivers to a user watching a soccer game a personalized version of the real-time streamed game. We develop a system integrating knowledge about the user, real time facial expression analysis and real time soccer game analyser in order to predict user’s intention on the soccer content delivery method. Several processes are required for this system: face and emotion recognition, soccer game analysis, soccer game prediction and intention estimation. In this talk I will present the current advancement in the development of the Live-Feeling Communication platform. In particular I will show the result in the soccer game analysis, prediction and emotional expressions recognition related to regional diversity. The main results of the current state of the art in both the emotion from facial expressions and soccer game analysis will also be presented and compared with the inhouse developed approaches.
