Learning from Dependent Data

Zimin, Alexander (2018) Learning from Dependent Data. PhD thesis, IST Austria.

[img] Text
Thesis.pdf - Published Version
Available under License All rights reserved.
Download (1011Kb)
[img] Archive
Restricted to Repository staff only
Available under License All rights reserved.
Download (622Kb)


The most common assumption made in statistical learning theory is the assumption of the independent and identically distributed (i.i.d.) data. While being very convenient mathematically, it is often very clearly violated in practice. This disparity between the machine learning theory and applications underlies a growing demand in the development of algorithms that learn from dependent data and theory that can provide generalization guarantees similar to the independent situations. This thesis is dedicated to two variants of dependencies that can arise in practice. One is a dependence on the level of samples in a single learning task. Another dependency type arises in the multi-task setting when the tasks are dependent on each other even though the data for them can be i.i.d. In both cases we model the data (samples or tasks) as stochastic processes and introduce new algorithms for both settings that take into account and exploit the resulting dependencies. We prove the theoretical guarantees on the performance of the introduced algorithms under different evaluation criteria and, in addition, we compliment the theoretical study by the empirical one, where we evaluate some of the algorithms on two real world datasets to highlight their practical applicability.

Item Type: Thesis (PhD)
DOI: 10.15479/AT:ISTA:TH1048
Subjects: 000 Computer science, knowledge & general works > 000 Computer science, knowledge & systems > 004 Data processing & computer science
500 Science > 510 Mathematics > 519 Probabilities & applied mathematics
Research Group: Lampert Group
Depositing User: Alexander Zimin
Date Deposited: 26 Sep 2018 11:36
Last Modified: 27 Sep 2018 13:46
URI: https://repository.ist.ac.at/id/eprint/1048

Actions (login required)

View Item View Item