2026-09-22T11:07:19Z https://www.ideals.illinois.edu/oai-pmh

oai:www.ideals.illinois.edu:2142/122271 2024-03-02 col_2142_5131 col_2142_8888 com_2142_5130 com_2142_8887 com_2142_234

Zhao, Han Li, Bo 2023-12 application/pdf en text Submission published under a 24 month embargo labeled 'Closed Access', the embargo will last until 2025-12-01 The student, Haoxiang Wang, accepted the attached license on 2023-12-06 at 15:20. The student, Haoxiang Wang, submitted this Thesis for approval on 2023-12-06 at 15:30. This Thesis was approved for publication on 2023-12-07 at 16:24. DSpace SAF Submission Ingestion Package generated from Vireo submission #20157 on 2024-03-01 at 13:57:41 Understanding gradual domain adaptation: Improved analysis, optimal path and beyond Wang, Haoxiang Machine Learning Domain Adaptation 2023-12-07 The vast majority of existing algorithms for unsupervised domain adaptation (UDA) focus on adapting from a labeled source domain to an unlabeled target domain directly in a one-off way. Gradual domain adaptation (GDA), on the other hand, assumes a path of $(T-1)$ unlabeled intermediate domains bridging the source and target, and aims to provide better generalization in the target domain by leveraging the intermediate ones. Under certain assumptions, Kumar et al. (2020) proposed a simple algorithm, Gradual Self-Training, along with a generalization bound in the order of $e^{O(T)} \left(\varepsilon_0+O\left(\sqrt{log(T)/n}\right)\right)$ for the target domain error, where $\varepsilon_0$ is the source domain error and $n$ is the data size of each domain. Due to the exponential factor, this upper bound becomes vacuous when $T$ is only moderately large. In this work, we analyze gradual self-training under more general and relaxed assumptions, and prove a significantly improved generalization bound as $\widetilde{O}\left(\varepsilon_0 + T\Delta + T/\sqrt{n} + 1/\sqrt{nT}\right)$, where $\Delta$ is the average distributional distance between consecutive domains. Compared with the existing bound with an exponential dependency on $T$ as a multiplicative factor, our bound only depends on $T$ linearly and additively. Perhaps more interestingly, our result implies the existence of an optimal choice of $T$ that minimizes the generalization error, and it also naturally suggests an optimal way to construct the path of intermediate domains so as to minimize the accumulative path length $T\Delta$ between the source and target. To corroborate the implications of our theory, we examine gradual self-training on multiple semi-synthetic and real datasets, which confirms our findings. We believe our insights provide a path forward toward the design of future GDA algorithms. Text eng https://hdl.handle.net/2142/122271 Copyright 2023 Haoxiang Wang M.S. Thesis Electrical & Computer Engr University of Illinois at Urbana-Champaign Electrical & Computer Eng