Forskning om prestandan i maskininlärning under beräkningsöverväganden
We are interested in finding answers to the open questions including: a) How to predict the best achievable performance for a given ML task and a given data set, under a limited amount of computation. b) How to predict the performance of the existing methods, as compared to the best achievable performance. c) How to achieve the best performance by modifying training specifications or developing new algorithm. As part of this research, areas of application will be targeted, where the above-mentioned questions are of paramount significance. This includes ML tasks under uncertainty or time variability.
PhD recruitment is in process.
Denna sidan finns endast på svenska
- Chalmers AI-forskningscentrum (CHAIR) (Centrumbildning, Sweden)