Algorithm Selection and Meta Learning
The Meta-Learning is one of the available approaches for optimization of machine learning, knowledge discovery and reasoning. It offers a variety of possibilities such as algorithm design, algorithm merging, algorithm selection etc. In our group we study the application of algorithm selection and meta learning to real-world problems. In particular:
-
Optimization of computer vision
-
Massive Ensembles for Algorithm Selection
-
Using alternative sources of information for meta-learning
-
Reasoning for algorithm selection
Publications
- Lukac, M. and Abdiyeva, K. and Bayanov, A. and Li, Albina, and Izbassarova, N. and Gabidolla, M. and Kameyama, M., Selecting Agorithms without meta-features, In: Del Bimbo A. et al. (eds) Pattern Recognition. ICPR International Workshops and Challenges. ICPR 2021. Lecture Notes in Computer Science, vol 12664. Springer, Cham. https://doi.org/10.1007/978-3-030-68799-1_44
- Lukac M., Izbassarova N., Li A., Kameyama M., Algorthm Selection for Non-Linearly Separable Algorithms in Computer Vision, Euromicro Workshop on ML-IoT, Online, 2018
- Lukac M., Abdiyeva K., Kim A., Kameyama M., Reasoning and Algorithm Selection Augmented Symbolic Segmentation, IEEE Technically Sponsored Intelligent Systems Conference (IntelliSys), pp. 259-266, 2017.
- Lukac M., Abdiyeva K., Kameyama M., Evaluation of Component Algorithms in an Algorithm Selection Approach to Semantic Segmentation Based on High-Level Information Feedback, Radio Electronics, Computer Science, Control, No. 1, 2016, pp. 92-100, DOI 10.15588/1607-3274-2016-1-11
- Lukac M., Kameyama M., Bayesian-Network-Based Algorithm Selection with High Level Representation Feedback for Real-World Information Processing, IT in Industry, Vol. 3, Issue. 1, pp.10-15, May 2015
- Lukac, M. & Kameyama, M. Int. J. Mach. Learn. & Cyber. (2015) 6: 417. https://doi.org/10.1007/s13042-013-0197-x