Connectionists: PhD position on neural networks for affective computing

[Marcel van Gerven]

PhD candidate: machine learning approaches for affective computing in humans and robots (1.0 FTE)

The goal of this PhD project is to develop neural network architectures that implement abstract cognitive functions, such as predicting someone’s emotional state or personality. Recent research has shown that this can be achieved using deep neural networks that predict such states using audio-visual input (arxiv.org/abs/1609.05119). In this project, you will further develop this framework for understanding affective computing in humans and endowing humanoid robots with these capabilities. 
  
The first part of this project aims to develop new explainable neural networks that allow us to understand how humans make decisions in high-level affective states. The goal here is to implement existing approaches and develop new approaches that improve our understanding of how neural networks arrive at their conclusions. Outcomes will be compared with behavioural data obtained for humans that perform the same task.
  
The second part of this project aims to develop efficient deep learning architectures that run on mobile devices in order to allow humanoid robots (www.ald.softbankrobotics.com/en/cool-robots/pepper) to interact more effectively with humans. The goal here is to develop and implement new approaches for creating models with low memory and low energy consumption that nevertheless perform optimally.

Work environment

The Donders Institute for Brain, Cognition and Behaviour focuses on state-of-the-art cognitive neuroscience using a multidisciplinary approach and offers excellent lab and neuroimaging facilities, PhD supervision, courses and technical support. The project will be embedded in the Donders research themes: Perception, Action and Control as well as Brain Networks and Neuronal Communication. 
  
The project will be embedded at the Centre for Cognition of the Donders Institute, which is also part of the Faculty of Social Sciences, one of the largest faculties at Radboud University in Nijmegen. As a PhD student you will be an active member of the research groups run by Dr Marcel van Gerven (www.ccnlab.net) and Dr Pim Haselager (www.socsci.ru.nl/haselag/). As a researcher you are expected to contribute to the neural networks and cognitive robotics lines within the Donders Institute’s AI department.

What we expect from you

We are looking for a technically skilled PhD candidate with a background in artificial intelligence, computer science, applied mathematics or a related field of research. You should be driven to push forward deep learning theories and be interested in cognitive psychology and robotics applications. 
  
You should have excellent mathematical skills, be a skilled Python programmer and have demonstrable experience with the development of deep learning algorithms. You should be aware of new developments in the deep learning field and be able to implement these developments. 
  
The ideal candidate is highly motivated, technically skilled, a team player and driven to push forward the development of the next generation of neural-network based AI systems. Your written and spoken English skills should be excellent. As a PhD student, you will devote a small part of your time to teaching in the AI department.

Please note that all PhD students at the Donders Centre for Cognition have a 0.9 FTE research appointment and a 0.1 FTE teaching appointment. You will be part of the Donders Graduate School for Cognitive Neuroscience (www.ru.nl/donders/graduate-school/).

For more information about this vacancy, please contact: 

Marcel van Gerven, Associate Professor, Principal investigator
Telephone: +31 24 3655931
E-mail: m.vangerven at donders.ru.nl

Pim Haselager
Telephone: +31 24 3616148
E-mail: w.haselager at donders.ru.nl

Are you interested?
You should upload your application (attn. of Marcel van Gerven) exclusively using the button 'Apply’ via the following link:

http://www.ru.nl/werken/details/details_vacature_0/?recid=597320

Your application should include (and be limited to) the following attachment(s):
	• application letter
	• CV (including contact information for at least two references)
	• example of written material (e.g. MSc thesis, project report) and proof of programming experience (if applicable)