MASTER THESIS
Supervisor: prof. dr. Frank van der Velde
35EC
Modeling and implementing aspects of cognitive processing forms the basis of many new applications in ICT systems, as used in society ("smart society") and industry ("smart industry"). An example is the success of "Deep Learning" (DL) in ICT systems such as self-driving cars, face recognition, speech recognition, and other forms of pattern recognition. DL models now begin to outperform humans on certain tasks. This offers possibilities for many new applications (e.g., in medical data classification).
The extent and importance of these applications will continue and increase in the near future. HFE students will most likely be directly confronted with these applications in their professional career. Hence, HFE students would profit by combining their skills in human factors with a basic understanding of and experience with these applications.
Recent developments in computer modeling have resulted in modeling tools that can be used by HFE students to model aspects of cognitive processing and implement them in systems such as robots. Some computer training is required, such as a basic knowledge of Python as taught in the HFE module. But the use of these tools does not require a profound experience with computer modeling.
In these master thesis projects a choice can be made on the task that will be modeled. This could be a (neural network) model of motor behavior, pattern classification, or concept development in language. In each of these domains, learning can be modeled using tools such as DL. The models developed can be implemented on a robot such as the iCub robot at the university of Twente.
Interested?
Contact: prof.dr. Frank van der Velde: f.vandervelde@utwente.nl
SUPERVISORS: PROF. DR. FRANK VAN DER VELDE, DR. FARIBA SHARIFIAN (IFADO)
35EC
In this project, we utilize computational and data-mining techniques to model cognitive processes and their modulations by different factors. We plan to use electroencephalogram (EEG) signals together with behavioral data (e.g. response times and performances) to build mathematical models of neural mechanism underlying basic cognitive processes. Such models provide powerful tools to compare normal versus abnormal neural interactions, as well as how they would be affected by different factors such as stress, age, and medical conditions.
Specifically, the assignment is focused on basic neurocognitive experiments such as visual attention, vigilance, executive functions, statistical learning, and speech perception. The project includes design of the selected experiments, as well as, collection of multi-dimensional data like behavioral, demographical, biomedical and neuro-physiological measures. We employ data-driven approaches to model the collected neural data, and expect these models to provide additional information to existing knowledge from theory-driven studies. Students have the opportunity to expand their knowledge about behavioral and EEG data collections, statistical /computational analysis, and scientific writing. The project definition is general and details of the computational techniques, and the cognitive tasks could be discussed.