Head of laboratory:Olzhas Makhambetov
The laboratory of computer science conducts research in the areas of computer science and information technology to provide solutions to the most challenging and pressing problems of Kazakhstan and the world.
Data Analysis and Visualization
Big Data Analysis and Visualization group conducts research and development of methods and techniques on the intersection of several Computer Science areas, including, but not limited to, topological data analysis, data mining, machine learning, visualization, numerical methods, wavelet analysis etc. Since a fair share of their research work is theoretic, a significant portion of the outcomes becomes applicable in a wide range of settings, thus finding their utility in numerous application domains (chemical data analysis, HPC/performance data analysis, analysis and mining massive text data sets, image recognition, visual analytics, psychometrics, etc.).
One of the major research undertakings of the group is the research and development of cloud-based visual analytics technologies, with the focus on two main types of large-scale data: multidimensional numerical data and massive text data. The aforementioned technologies are due for an eventual integration with a central cloud platform, which would include a variety of solutions by other groups as well.
Cloud and High-Performance Computing
For the efficient use of existing computing resources the group is deploying cloud infrastructure platform based on OpenStack. This platform allows to combine all existing physical machines into a single pool of virtual resources, while providing a simple interface to users (research groups), and administrators as well.
Currently deployed Openstack Kilo is capable to pass through graphics cards (NVidia Tesla K20) to virtual machines for parallel computing on CUDA. Further development of the system includes the flexible user management, enhanced security and scalability (fault tolerance, high availability), as well as creating an environment for high-performance and distributed computing (Apache Spark, MPI, NoSQL databases).
The computational linguistics research group is working on the development of natural language processing methods and algorithms, putting emphasis on studying the problems of machine translation, semantic and social network analysis, recommender systems, etc.
Signal and speech processing group
Our group is focused on developing methods and algorithms for Kazakh speech recognition and synthesis using machine learning and artificial intelligence approaches. Also we do research in general signal processing, acoustics and phonetics. The aim of the group is to develop the useful and friendly interface for human computer interaction, information search and retrieval via speech, the most natural means of communication for human. Although Kazakh is the primary object of our research, we exploit language independent approaches for speech processing. Additionally, we pay special attention to collecting and processing speech corpora for a wide range of users.
The multimedia technologies group goes beyond established borders and disciplines. In their research the group encourages unconventional approaches by intermingling and juxtaposition of seemingly unrelated research areas. It gives an opportunity to create innovative technologies centered around an intersection of a diversity of research fields, such as computer vision, computational linguistics, logopedics, pedagogy, psychology etc. Nowadays, research group has been working on the sign language recognition system development for the hearing impaired people, smart classroom technologies and interactive marketing technologies.