國立高雄大學圖資館 |

登入

回首頁

On Some Problems in Transfer Learning /

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	On Some Problems in Transfer Learning /Nicholas R Galbraith.
作者:	Galbraith, Nicholas R.,
面頁冊數:	1 electronic resource (121 pages)
附註:	Source: Dissertations Abstracts International, Volume: 85-08, Section: B.
附註:	Advisors: Kpotufe, Samory K.; Namkoong, Hongseok Committee members: Maleki, Arian; Sen, Bodhisattva; Zheng, Tian.
Contained By:	Dissertations Abstracts International85-08B.
標題:	Statistics.
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30817247
ISBN:	9798381681031

On Some Problems in Transfer Learning /
Galbraith, Nicholas R.,

On Some Problems in Transfer Learning /Nicholas R Galbraith. - 1 electronic resource (121 pages)

Source: Dissertations Abstracts International, Volume: 85-08, Section: B.

This thesis consists of studies of two important problems in transfer learning: binary classification under covariate-shift transfer, and off-policy evaluation in reinforcement learning. First, the problem of binary classification under covariate shift is considered, for which the first efficient procedure for optimal pruning of a dyadic classification tree is presented, where optimality is derived with respect to a notion of average discrepancy between the shifted marginal distributions of source and target. Further, it is demonstrated that the procedure is adaptive to the discrepancy between marginal distributions in a neighbourhood of the decision boundary. It is shown how this notion of average discrepancy can be viewed as a measure of relative dimension between distributions, as it relates to existing notions of information such as the Minkowski and Renyi dimensions. Experiments are carried out on real data to verify the efficacy of the pruning procedure as compared to other baseline methods for pruning under transfer.The problem of off-policy evaluation for reinforcement learning is then considered, where two minimax lower bounds for the mean-square error of off-policy evaluation under Markov decision processes are derived. The first of these gives a non-asymptotic lower bound for OPE in finite state and action spaces over a model in which the mean reward is perturbed arbitrarily (up to a given magnitude) that depends on an average weighted chi-square divergence between the behaviour and target policies. The second provides an asymptotic lower bound for OPE in continuous state-space when the mean reward and policy ratio functions lie in a certain smoothness class. Finally, the results of a study that purported to have derived a policy for sepsis treatment in ICUs are replicated and shown to suffer from excessive variance and therefore to be unreliable; our lower bound is computed and used as evidence that reliable off-policy estimation from this data would have required a great deal more samples than were available.

English

ISBN: 9798381681031Subjects--Topical Terms:

182057
Statistics.
Subjects--Index Terms:

Decision trees

On Some Problems in Transfer Learning /
LDR:03539nmm a22004333i 4500 001 676277
005 20250420214332.5
006 m o d
007 cr|nu||||||||
008 250901s2024 miu||||||m |||||||eng d
020 $a 9798381681031
035 $a (MiAaPQD)AAI30817247
035 $a AAI30817247
040 $a MiAaPQD $b eng $c MiAaPQD $e rda
100 1 $a Galbraith, Nicholas R., $e author. $3 990022
245 1 0 $a On Some Problems in Transfer Learning / $c Nicholas R Galbraith.
264 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2024
300 $a 1 electronic resource (121 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 85-08, Section: B.
500 $a Advisors: Kpotufe, Samory K.; Namkoong, Hongseok Committee members: Maleki, Arian; Sen, Bodhisattva; Zheng, Tian.
502 $b Ph.D. $c Columbia University $d 2024.
520 $a This thesis consists of studies of two important problems in transfer learning: binary classification under covariate-shift transfer, and off-policy evaluation in reinforcement learning. First, the problem of binary classification under covariate shift is considered, for which the first efficient procedure for optimal pruning of a dyadic classification tree is presented, where optimality is derived with respect to a notion of average discrepancy between the shifted marginal distributions of source and target. Further, it is demonstrated that the procedure is adaptive to the discrepancy between marginal distributions in a neighbourhood of the decision boundary. It is shown how this notion of average discrepancy can be viewed as a measure of relative dimension between distributions, as it relates to existing notions of information such as the Minkowski and Renyi dimensions. Experiments are carried out on real data to verify the efficacy of the pruning procedure as compared to other baseline methods for pruning under transfer.The problem of off-policy evaluation for reinforcement learning is then considered, where two minimax lower bounds for the mean-square error of off-policy evaluation under Markov decision processes are derived. The first of these gives a non-asymptotic lower bound for OPE in finite state and action spaces over a model in which the mean reward is perturbed arbitrarily (up to a given magnitude) that depends on an average weighted chi-square divergence between the behaviour and target policies. The second provides an asymptotic lower bound for OPE in continuous state-space when the mean reward and policy ratio functions lie in a certain smoothness class. Finally, the results of a study that purported to have derived a policy for sepsis treatment in ICUs are replicated and shown to suffer from excessive variance and therefore to be unreliable; our lower bound is computed and used as evidence that reliable off-policy estimation from this data would have required a great deal more samples than were available.
546 $a English
590 $a School code: 0054
650 4 $a Statistics. $3 182057
650 4 $a Computer engineering. $3 212944
653 $a Decision trees
653 $a Nonparametrics
653 $a Evaluation
653 $a Pruning
653 $a Transfer learning
690 $a 0463
690 $a 0464
710 2 $a Columbia University. $b Statistics. $3 603132
720 1 $a Kpotufe, Samory K. $e degree supervisor.
720 1 $a Namkoong, Hongseok $e degree supervisor.
773 0 $t Dissertations Abstracts International $g 85-08B.
790 $a 0054
791 $a Ph.D.
792 $a 2024
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30817247