Umeå University, the Department of Computing Science, is seeking outstanding candidates for a postdoc position in robust machine learning and data-centric optimization with focus on scarce data and non-standard model settings.
The rapid increase of autonomous systems and applications are providing challenges in dealing with both scarce and petabytes of data in diverse environments. Machine learning has recently achieved a lot with the standard assumption that availability of data is large. However, it still remains many challenges in areas where this is not true. These sizes and heterogenous features make the machine learning models larger and more complex. Classical approaches to training, learning, and inference fail to address the problems of scarce data, non-standard model settings, data-centric optimization (centralized and distributed), communication, computation, synchronization, and many more. This project focuses on design and implementation of robust machine learning and data-centric optimization algorithms for scarce and petabytes of data and non-standard model settings, which are ideally suited for constraint environments and edge infrastructures.
This project aims to design and implement robust learning and data-centric optimization techniques for advancing state-of-the-art machine learning algorithms where data is geographically distributed, sensitive, and scarce. Robust machine learning and data-centric optimization algorithms empower models through multi-level (local, global and hybrid) training, learning, and inference with data-centric optimization for scarce data and non-standard model settings. By creating unique features (e.g., decentralized training, learning and inference, fault-tolerant against failures and attacks, data-centric optimization, robustness), this project addresses the challenges in the following areas: robust learning; learning with scarce data and non-standard model settings; lack of theoretical knowledge to build manual models; computation efficient learning and optimization for obtaining more accurate and robust models with applications to constraint environments (i.e., Industrial Internet of Things (IIoT), healthcare systems) and edge infrastructures.