Scott Decatur. Learning in hybrid noise environments using statistical queries. In D. Fisher and H. J. Lenz, editors, Learning from Data: Artificial Intelligence and Statistics V. Springer Verlag,  Home/Search   Document Details and Download   Summary   Related Articles   Check

.... happens with probability j, the classification of the example is flipped and so the algorithm is provided with the, wrongly classified, example (x; 1 Gamma c t (x) Another (stronger) model, called the Malicious Noise model, was introduced in [23] revisited in [17] and was further studied in [8, 10, 11, 20]. In this model the adversary, whenever the j biased coin shows H , can replace the example (x; c t (x) by some arbitrary pair (x 0 ; b) where x 0 is any point in the input space and b is a boolean value. Note that this in particular gives the adversary the power to distort the ....

.... Noise Nasty Classification Noise Point and Label Noise Malicious Noise Nasty Sample Noise Table 1: Summary of models for PAC learning from noisy data We argue that the newly introduced model, not only generalizes the previous noise models, including variants such as Decatur s CAM model [11] and CPCN model [12] but also, that in many real world situations, the assumptions previous models made about the noise seem unjustified. For example, when training data is the result of some physical experiment, noise may tend to be stronger in boundary areas rather than being uniformly ....

S. E. Decatur, "Learning in Hybrid Noise Environments Using Statistical Queries", in Learning from Data: Artificial Intelligence and Statistics V, D. Fisher and H. J. Lenz, (Eds.), 1996.

....and its data, and has the property that any algorithm for learning in the SQ model can automatically be converted to an algorithm for learning in the presence of random classification noise in the standard PAC model. This result has been extended to more general forms of noise as well [Dec93, Dec96] The importance of the Statistical Query model is attested to by the fact that before its introduction, there were only a few provably noise tolerant learning algorithms, whereas now it is recognized that almost all known learning algorithms can be formulated as SQ algorithms, and hence can be ....

S. E. Decatur. Learning in hybrid noise environments using statistical queries. In D. Fisher and H.-J. Lenz, editors, Learning from Data: Artificial Intelligence and Statistics V. Springer Verlag, 1996.

....from the property that statistical queries can also be simulated with the use of noisy example oracles. Specifically, an SQ algorithm can be simulated in the PAC model in the presence of classification noise, malicious errors, attribute noise and even hybrid models combining these different noises [14, 6, 8, 7]. A key parameter in the complexity of the PAC algorithm generated by the simulation of SQ algorithms is the tolerance of the SQ algorithm, which quantifies the largest additive error that the SQ algorithm can tolerate when receiving an answer to its most sensitive query. The limitation of the ....

Scott Decatur. Learning in hybrid noise environments using statistical queries. In D. Fisher and H. J. Lenz, editors, Learning from Data: Artificial Intelligence and Statistics V. Springer Verlag,

....the property that statistical queries can actually be simulated with the use of noisy example oracles. Specifically, an SQ algorithm can be simulated in the PAC model in the presence of classification noise, malicious errors, attribute noise and even hybrid models combining these different noises [12, 5, 7, 6]. A key parameter in the complexity of the PAC algorithm generated by the simulation of SQ algorithms is the tolerance of the SQ algorithm, which quantifies the largest additive error that the SQ algorithm can tolerate when receiving an answer to its most sensitive query. The limitation of ....

Scott Decatur. Learning in hybrid noise environments using statistical queries. In Proceedings of the Fifth International Workshop on Artificial Intelligence and Statistics, pages 175--185, January 1995. To appear in Springer-Verlag Lecture Notes in Statistics.

....Bibliographical Notes Most of the results of this thesis have been published previously. The results contained in Chapters 3 and 5, as well as some of Chapter 6, appear in a Harvard University Technical Report HUTR 17 94 (Aslam and Decatur, 1994) and an extended abstract in COLT 95 (Aslam and Decatur, 1995). The remainder of Chapter 6 and all of Chapter 8 and Appendix B appear as an extended abstract in COLT 93 (Decatur, 1993) An extended abstract of the results in Chapter 4 and Appendix A appear in FOCS 93 (Aslam and Decatur, 1993) Most of the results of Chapter 9 appear as an extended abstract ....

.... abstract in COLT 93 (Decatur, 1993) An extended abstract of the results in Chapter 4 and Appendix A appear in FOCS 93 (Aslam and Decatur, 1993) Most of the results of Chapter 9 appear as an extended abstract in the proceedings of AI STATS 95 and SpringerVerlag Lecture Notes in Statistics (Decatur, 1995). The remainder of Chapter 9 and all of Chapter 7 appear as an extended abstract in COLT 95 (Decatur and Gennaro, 1995) ix C h a p t e r 1 Introduction in duc tion (in duk 0 shn) n. 3.a. Logic. A principle of reasoning to a conclusion about all the members of a class from examination of only ....

Decatur, Scott. (1995). Learning in hybrid noise environments using statistical queries. In Proceedings of the Fifth International Workshop on Artificial Intelligence and Statistics, pages 175--185, January. To appear in Springer-Verlag Lecture Notes in Statistics.

....stems from the property that statistical queries can actually be simulated with the use of noisy example oracles. Specifically, an SQ algorithm can be simulated in the PAC model in the presence of classification noise, malicious errors, and even hybrid models combining these two types of noise [8, 5, 6]. In order to relate statistical query learning to PAC learning with attribute noise, we define a new complexity measure on statistical query algorithms called the view. The view of an SQ algorithm is the maximum over all queries in the algorithm, of the number of input variables on which the ....

Scott Decatur. Learning in hybrid noise environments using statistical queries. In Proceedings of the Fifth International Workshop on Artificial Intelligence and Statistics, pages 175--185, January 1995. To appear in Springer-Verlag Lecture Notes in Statistics.

No context found.

S. E. Decatur. Learning in hybrid noise environments using statistical queries. In D. Fisher and H.-J. Lenz, editors, Learning from Data: Artificial Intelligence and Statistics V. Springer Verlag, 1996.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC