Algorithms for Hyper-Parameter Optimization
- J. Bergstra, R. Bardenet, Yoshua Bengio, B. Kégl
- 12 December 2011
Computer Science
NIPS
This work contributes novel techniques for making response surface models P(y|x) in which many elements of hyper-parameter assignment (x) are known to be irrelevant given particular values of other elements.
Learning and Design of Principal Curves
- B. Kégl, A. Krzyżak, T. Linder, K. Zeger
- 1 March 2000
Computer Science
IEEE Transactions on Pattern Analysis and Machine…
This work defines principal curves as continuous curves of a given length which minimize the expected squared distance between the curve and points of the space randomly chosen according to a given distribution, making it possible to theoretically analyze principal curve learning from training data and it also leads to a new practical construction.
Collaborative hyperparameter tuning
- R. Bardenet, M. Brendel, B. Kégl, M. Sebag
- 16 June 2013
Computer Science
International Conference on Machine Learning
A generic method to incorporate knowledge from previous experiments when simultaneously tuning a learning algorithm on new problems at hand is proposed and is demonstrated in two experiments where it outperforms standard tuning techniques and single-problem surrogate-based optimization.
Correlation of the highest-energy cosmic rays with nearby extragalactic objects.
- J. Abraham, P. Abreu, B. Kégl
- 9 November 2007
Physics
Science
Using data collected at the Pierre Auger Observatory during the past 3.7 years, we demonstrated a correlation between the arrival directions of cosmic rays with energy above 6 x 10(19) electron volts…
Aggregate features and ADABOOST for music classification
- J. Bergstra, Norman Casagrande, D. Erhan, D. Eck, B. Kégl
- 1 December 2006
Computer Science
Machine-mediated learning
An algorithm that predicts musical genre and artist from an audio waveform using the ensemble learner ADABOOST and evidence collected from a variety of popular features and classifiers that the technique of classifying features aggregated over segments of audio is better than classifying either entire songs or individual short-timescale features.
The Pierre Auger Cosmic Ray Observatory
- A. Aab, P. Abreu, F. Zuccarello
- 1 June 1998
Physics
Piecewise linear skeletonization using principal curves
- B. Kégl, A. Krzyżak
- 3 September 2000
Computer Science
Proceedings 15th International Conference on…
The algorithm is an extension of the polygonal line algorithm, originally designed to find principal curves of data sets, to compute the principal graph of a data set.
Correlation of the highest-energy cosmic rays with the positions of nearby active galactic nuclei
- J. Abraham, P. Abreu, M. Ziolkowski
- 1 April 2008
Physics
Observation of the suppression of the flux of cosmic rays above 4 x 10 (19) eV.
- J. Abraham, P. Abreu, M. Ziolkowski
- 8 August 2008
Physics
Physical Review Letters
The energy spectrum of cosmic rays above 2.5 x 10;{18} eV, derived from 20,000 events recorded at the Pierre Auger Observatory, is described and the hypothesis of a single power law is rejected with a significance greater than 6 standard deviations.
SEARCHES FOR ANISOTROPIES IN THE ARRIVAL DIRECTIONS OF THE HIGHEST ENERGY COSMIC RAYS DETECTED BY THE PIERRE AUGER OBSERVATORY
- A. Aab, P. Abreu, F. Zuccarello
- 1 May 2015
Physics
We analyze the distribution of arrival directions of ultra-high-energy cosmic rays recorded at the Pierre Auger Observatory in 10 years of operation. The data set, about three times larger than that…
...
...