Optimization algorithm design
To deal with changing and uncertain problems, our research includes dynamic algorithm selection and its configuration. This is achieved through problem landscape analysis and supported by comparative robust statistics. Our approaches can handle either single objective or multiple and contradictory objectives.
To process large amounts of data with artificial intelligence and computational intelligence approaches, we research different machine learning techniques, representational learning techniques as well as natural language processing. These are also supported by deep neural networks.
Data management and visualization
To better manage and visualize any data, our research includes semantic web and knowledge graph. Several human-computer interaction aspects are considered when getting and sending data. The interaction interfaces are also studied through usability testing.
Adaptive computing platforms
For faster and more reliable execution of algorithms, FPGA and GPGPU processing acceleration are used, along with the research of the dynamically reconfigurable FPGAs. The approaches of approximate computing are also very welcome in the algorithm's implementation as edge computing.