Stability and Hypothesis Transfer Learning
نویسندگان
چکیده
We consider the transfer learning scenario, where the learner does not have access to the source domain directly, but rather operates on the basis of hypotheses induced from it – the Hypothesis Transfer Learning (HTL) problem. Particularly, we conduct a theoretical analysis of HTL by considering the algorithmic stability of a class of HTL algorithms based on Regularized Least Squares with biased regularization. We show that the relatedness of source and target domains accelerates the convergence of the Leave-OneOut error to the generalization error, thus enabling the use of the Leave-One-Out error to find the optimal transfer parameters, even in the presence of a small training set. In case of unrelated domains we also suggest a theoretically principled way to prevent negative transfer, so that in the limit we recover the performance of the algorithm not using any knowledge from the source domain.
منابع مشابه
Cross-Linguistic Transfer or Target Language Proficiency: Writing Performance of Trilinguals vs. Bilinguals in Relation to the Interdependence Hypothesis
This study explored the nature of transfer among bilingual vs. trilinguals with varying levels of competence in English and their previous languages. The hypotheses were tested in writing tasks designed for 75 high (N= 35) vs. intermediate (N=40) proficient EFL learners with Turkish, Persian, English and Persian, English linguistic backgrounds. Qualitative data were also collected through some ...
متن کاملCorrection to “Stability and Hypothesis Transfer Learning”
There is an error in “Stability and Hypothesis Transfer Learning” (Kuzborskij & Orabona, 2013) which appeared in proceedings of ICML 2013. The Leave-One-Out generalization bound for Hypothesis Transfer Learning algorithm through Regularized Least Squares with biased regularization does not have the right convergence rate with respect to the regularization parameter λ and the source risk on the ...
متن کاملA Theoretical Analysis of Metric Hypothesis Transfer Learning
We consider the problem of transferring some a priori knowledge in the context of supervised metric learning approaches. While this setting has been successfully applied in some empirical contexts, no theoretical evidence exists to justify this approach. In this paper, we provide a theoretical justification based on the notion of algorithmic stability adapted to the regularized metric learning ...
متن کاملI-18: Techniques and Technologies for Embryo Transfer: Does It Really Matter.
The learning objectives of this presentation are to understand the dynamics involved in the process of ET, evaluate the evidence for/against common practices and techniques and develop a standardized ET process in view of supporting evidence. Gametes and embryos are handled with extreme care at every step of the Laboratory process. ET is the least sophisticated step of the in vitro Fertilizatio...
متن کاملThe production of lexical categories (VP) and functional categories (copula) at the initial stage of child L2 acquisition
This is a longitudinal case study of two Farsi-speaking children learning English: ‘Bernard’ and ‘Melissa’, who were 7;4 and 8;4 at the start of data collection. The research deals with the initial state and further development in the child second language (L2) acquisition of syntax regarding the presence or absence of copula as a functional category, as well as the role and degree of L1 influe...
متن کامل