Title :
Semi-supervised multi-domain regression with distinct training sets
Author :
Michaeli, Tomer ; Eldar, Yonina C. ; Sapiro, Guillermo
Author_Institution :
Technion-Israel Inst. of Technol., Haifa, Israel
Abstract :
We address the problems of multi-domain and single-domain regression based on distinct labeled training sets for each of the domains and a large unlabeled training set from all domains. We formulate these problems as ones of Bayesian estimation with partial knowledge of statistical relations. We propose a worst-case design strategy and study the resulting estimators. Our analysis explicitly accounts for the cardinality of the labeled sets and includes the special cases in which one of the labeled sets is very large or, in the other extreme, completely missing. We demonstrate our estimators in the context of audio-visual word recognition and provide comparisons to several recently proposed multi-modal learning algorithms.
Keywords :
Bayes methods; audio-visual systems; learning (artificial intelligence); regression analysis; speaker recognition; video signal processing; Bayesian estimation; audio-visual word recognition; distinct training sets; multimodal learning algorithms; semisupervised multidomain regression; single-domain regression; spoken digit classification; statistical relation partial knowledge; unlabeled training set; video feature-vector; worst-case design strategy; Accuracy; Bayesian methods; Estimation; Joints; Testing; Training; Visualization; Bayesian estimation; multi-modal learning;
Conference_Titel :
Acoustics, Speech and Signal Processing (ICASSP), 2012 IEEE International Conference on
Conference_Location :
Kyoto
Print_ISBN :
978-1-4673-0045-2
Electronic_ISBN :
1520-6149
DOI :
10.1109/ICASSP.2012.6288336