Seminaari: 58302301 Koneoppimisen teoria (2 ov)

Prof. Tapio Elomaa

Syksy 2002, 13.9.-29.11. pe 12-14 B450

Yleistä

Seminaarissa käsitellään uusia analyyttisesti motivoituja koneoppimismenetelmiä. Varsinkin tukivektorikoneita ja oppimisen tehostamista koskevia uusia tuloksia tarkastellaan.

Esitiedot

Kurssin Koneoppiminen tiedot ovat välttämätön edellytys osallistumiselle. Algoritmien suunnittelu ja analyysi. Riittävät matematiikan valmiudet.

Osallistujat

Autio Ilkka, Borrás Juan Carlos, Forsblom Ilkka, Haapasalo Jaakko, Koskenniemi Ilkka, Kääriäinen Matti, Lindgren J. T., Löfström Jaakko, Malinen Tuomo, Mielikäinen Taneli, Rantanen Ari, Rousu Juho

Aikataulu

13. 9.	Järjestäytyminen
20. 9.	FM Matti Kääriäinen: Satunnaisprojektiot yleistysvirheanalyysissä
	Ashutosh Garg, Sariel Har-Peled, Dan Roth: On generalization bounds, projection profile, and margin distribution. Proc. 19th Intl. Conf. on Machine Learning (pp. 171-178), 2002.
27. 9.	FM Taneli Mielikäinen: Proximal Vector Machines
	Glenn Fung, Olvi L. Mangasarian: Proximal support vector machine classifiers. Proc. 7th ACM SIGKDD Intl. Conf. on Knowledge Discovery and Data Mining (pp. 77-86), 2001 Deepak K. Agarwal: Shrinkage Estimator Generalizations of Proximal Support Vector Machines. Proc. 8th ACM SIGKDD Intl. Conf. on Knowledge Discovery and Data Mining (pp. 173-182), 2002
4. 10.	FT Juho Rousu: On Kernel Optimization and Learning
	O. Chapelle, V. Vapnik, O. Bousquet, S. Mukherjee: Choosing Multiple Parameters for Support Vector Machines. Machine Learning 46 (2002) 131-159 N. Cristianini, J. Shawe-Taylor: An Introduction to Support Vector Machines and Other Kernel-Based Learning Methods. 2000 N. Cristianini, J. Shawe-Taylor, A. Elisseeff, J. Kandola: On Kernel-Target Alignment. Advances in Neural Information Processing Systems 14, 2002 Gert Lanckriet, Nello Cristianini, Peter Bartlett, Laurent El Ghaoui, Michael I. Jordan: Learning the Kernel Matrix with Semi-Definite Programming. Proc. 19th Intl. Conf. on Machine Learning (pp. 323-330), 2002.
11. 10.	FM J. T. Lindgren: Stochastic Discrimination
	Eugene M. Kleinberg: Stochastic discrimination. Annals of Mathematics and Artificial Intelligence 1 (1990) 207-239 Eugene M. Kleinberg: An overtraining-resistant stochastic modeling method for pattern recognition. Annals of Statistics 24, 6 (1996) 2319-2349 Eugene M. Kleinberg: On the algorithmic implementation of stochastic discrimination. IEEE Transactions on Pattern Analysis and Machine Intelligence 22, 5 (2000) 473-490
18. 10.	FM Tuomo Malinen: Convex Upper Bounding of Classification Error
	Tong Zhang: Statistical Behavior and Consistency of Support Vector Machines, Boosting, and Beyond. Proc. 19th Intl. Conf. on Machine Learning (pp. 690-697), 2002.
25. 10.	Ei seminaaria (FDK/BRU seminaari, Gustavelund)
1. 11.	FM Ilkka Autio: FastSLAM (Simultaneous Localization and Mapping)
	Michael Montemerlo, Sebastian Thrun, Daphne Koller, Ben Wegbreit: FastSLAM: a factored solution to simultaneous mapping and localization problem. Proc. 18th Natl. Conf. on Artificial Intelligence (pp. 593-598), 2002
8. 11.	M.Sc. Juan Carlos Borrás: Invariant SVMs
	Abstract: An interesting pattern found in an image is also interesting whether it appears as such, or rotated or shifted a few pixels on the image. The term invariance refers to the degree of variation up to which a pattern my be presented without changing its "meaning". The presentation describes proposals made in the field of SVMs in order to cope with the problem above. Practical results refer to the best up-to-today classification results over the USPS digits database. Olivier Chapelle, Bernhard Schölkopf: Incorporating Invariances in nonlinear support vector machines. Advances in Neural Information Processing Systems 14, 2002 Dennis Decoste, Berhnhard Schölkopf: Training invariant support vector machines. Machine Learning 46 (2002) 161-190
15. 11.	Kello 11.15-
	Ilkka Koskenniemi: Prior Knowledge and Unlabeled Data in Ensemble Methods
	Robert E. Schapire, Marie Rochery, Mazin Rahim, Narenda Gupta: Incorporating prior knowledge into boosting. Proc. 19th Intl. Conf. on Machine Learning (pp. 538-545), 2002. Kristin P. Bennett, Ayhan Demiriz, Richard Maclin: Exploiting unlabeled data in ensemble methods. Proc. 8th ACM SIGKDD Intl. Conf. on Knowledge Discovery and Data Mining (pp. 289-296), 2002
+	Jaakko Löfström: Adventures in Version Space
22. 11.	Kello 11.15-
	Jaakko Haapasalo: Optimal Bin-Packing with Branch-and-Bound Algorithms
	Richard E. Korf: A new algorithm for optimal bin packing. Proc. 18th Natl. Conf. on Artificial Intelligence (pp. 731-736), 2002
+	Ilkka Forsblom: Occam's Razor
	Ming Li, John Tromp, Paul M.B. Vitanyi: Sharpening Occam's razor. Proc. 8th Intl. Conf. Computing and Combinatorics (pp. 411-419), 2002
29. 11.	Ei seminaaria (Jaak Vilo väittelee aiheesta "Pattern Discovery from Biosequences", Auditorio klo 12)
2. 12.	FDK tutkimusseminaari A217 14-16
	Dos. Jyrki Kivinen: Online Learning of Linear Classifiers
	Abstract: This talk surveys some basic techniques and recent results related to online learning. Our focus is on linear classification. The most familiar algorithm for this task is the perceptron. We explain the perceptron algorithm and its convergence proof as an instance of a generic method based on Bregman divergences. This leads to a more general algorithm known as the p-norm perceptron. We generalise the perceptron convergence theorem for the p-norm perceptron and to the non-separable case. We also apply regularisation, again based on Bregman divergences, to make the algorithm more robust against target movement.
5. 12.	FM Ari Rantanen: Diffusion Kernels (Huom.: torstai, 14.15-, B450)
	Jaz Kandola, John Shawe-Taylor, Nello Cristianini: On the application of diffusion kernel to text data. NeuroCOLT Technical Report NC-TR-02-122 R. I. Kondor, J. Lafferty: Diffusion kernels on graphs and other discrete input spaces. Proc. 19th Intl. Conf. on Machine Learning (pp. 315-322), 2002

Linkkejä

NIPS*2001 Online Proceedings

Boosting Research Site

Kernel Machines

elomaa@cs.helsinki.fi