Thesis supervisor: Balázs Csanád Csáji
Location of studies (in Hungarian): Institute for Computer Science and Control, Hungarian Academy of Sciences Abbreviation of location of studies: MTA
Description of the research topic:
Machine learning is one of the dynamically growing fields of computer science, whose results are widely applied in practice from the traditional IT related fields (e.g., filtering spams, routing, and recommendation systems) to economics, medical and engineering applications (e.g., predicting financial time series, diagnoses of diseases, intelligent manufacturing).
Uncertainty plays a key role in many machine learning problems. It can have various sources, such as measurement or human errors, deterioration of components, lack of information, imprecise models, limited resources, environmental changes, and the inherent stochastic nature of some systems.
Such uncertainties call for statistical approaches. The research aims at theoretically and empirically analyzing and improving statistical machine learning methods, and exploring new directions.
Some fundamental questions are as follows: how can we learn (static or dynamic) models from (noisy, incomplete) experimental data; how can we measure the uncertainty of our models; how can we detect if the underlying system has changed; how can the model be recursively updated based on new data; how can we forecast the short-term behavior of the system; how can we exploit the model for efficient decision making; how can we give strong theoretical guarantees for the behavior of our learning algorithm based on minimal statistical assumptions; how can we efficiently explore the underlying system; how can we balance exploration and exploitation; and how can these learning tasks be distributed.
Required language skills: English Further requirements: probability theory basics, recommended: Matlab and/or R skills
Number of students who can be accepted: 2
Deadline for application: 2017-05-31
2024. IV. 17. ODT ülés Az ODT következő ülésére 2024. június 14-én, pénteken 10.00 órakor kerül sor a Semmelweis Egyetem Szenátusi termében (Bp. Üllői út 26. I. emelet).
Login
