History
The heuristic approach of self-training (also known as self-learning or self-labeling) is historically the oldest approach to semi-supervised learning, with examples of applications starting in the 1960s (see for instance Scudder (1965)).
The transductive learning framework was formally introduced by Vladimir Vapnik in the 1970s. Interest in inductive learning using generative models also began in the 1970s. A probably approximately correct learning bound for semi-supervised learning of a Gaussian mixture was demonstrated by Ratsaby and Venkatesh in 1995
Semi-supervised learning has recently become more popular and practically relevant due to the variety of problems for which vast quantities of unlabeled data are available—e.g. text on websites, protein sequences, or images. For a review of recent work see a survey article by Zhu (2008).
Read more about this topic: Semi-supervised Learning
Famous quotes containing the word history:
“Certainly there is not the fight recorded in Concord history, at least, if in the history of America, that will bear a moment’s comparison with this, whether for the numbers engaged in it, or for the patriotism and heroism displayed.”
—Henry David Thoreau (1817–1862)
“There is nothing truer than myth: history, in its attempt to “realize” myth, distorts it, stops halfway; when history claims to have “succeeded” this is nothing but humbug and mystification. Everything we dream is “realizable.” Reality does not have to be: it is simply what it is.”
—Eugène Ionesco (b. 1912)
“The history of work has been, in part, the history of the worker’s body. Production depended on what the body could accomplish with strength and skill. Techniques that improve output have been driven by a general desire to decrease the pain of labor as well as by employers’ intentions to escape dependency upon that knowledge which only the sentient laboring body could provide.”
—Shoshana Zuboff (b. 1951)