Classifiers That Improve With Use
1,834 views · Published 9 October 2007 · 48:06 · Indexed 8 October 2026
Channel: Google TechTalks · 2007 · Howto & Style
Google Tech Talks Januaary 29, 2007 ABSTRACT Training on imperfectly representative data inevitably leads to classification errors. Retraining an OCR engine with post-edited data, or even with the imperfect labels assigned by the classifier, reduces both bias and variance. Although the theoretical foundations of decision-directed adaptation are meager, it has proved successful in diverse experiments. When the operational data can be partitioned into isogenous subsets, style-constrained classification is appropriate. Patterns should be recognized in groups rather than in isolation. Shape and language context are complementary. Operator interaction should be rationalized. Only dynamic classifiers...
More from this channel
-
34:17
Thinglink - A Free Product Code for Creative Work
-
32:48
Thinking Beyond Borders
-
44:29
Understanding Urban Environments Through the Use of...
-
46:04
Customizable Scalable Compute Intensive Stream Queries
-
51:26
All The Government's Information
-
41:35
Combining Discriminative Features to Infer Complex...
-
59:42
Dimensions of Reputation in Electronic Markets
-
46:50
Accessing Legacy Documents in the iPod Age