Main

Home Page

I am Associate Researcher at the Computational Sciences Department of the National Institute of Astrophysics, Optics and Electronics (INAOE). See the Research section for more information about my background and interests.


News highlights:

(see the News section for past news):


New paper on Language and Vision published in Neurocomputing

A new paper was accepted and published online in the Neurocomputing Journal:

Improving the BoVW via Discriminative Visual N-Grams and MKL Strategies. A. Pastor López-Monroy, Manuel Montes-y-Gómez, Hugo Jair Escalante, Angel Cruz-Roa, Fabio A. González. Neurocomputing, available online: November 10, 2015. ( Impact factor: 2.083 -JCR)


This article by Adrián Pastor López reports research on the use of text mining techniques in image classification via the Bag-of-Words and Bag-of-Features analogy.



Paper published and presented in BMVC

A paper was accepted and presented at the British Conference on Computer Vision:

Gesture and Action Recognition by Evolved Dynamic Subgestures.. Víctor Ponce-López, Hugo Jair Escalante, Sergio Escalera and Xavier Baró. In Xianghua Xie, Mark W. Jones, and Gary K. L. Tam, editors, Proceedings of the British Machine Vision Conference (BMVC), pages 129.1-129.13. BMVA Press, September 2015..


This article by Victor Lopez describes a novel approach to learn subgestures by means of an evolutionary algorithm. The method is tested with DTW and HMM on benchmark data.



New paper about winning method in PAN2014 author profiling task

A new paper has been published online in the Journal of Knowledge-based Systems:

Discriminative Subprofile-Specific Representations for Author Profiling in Social Media. A. Pastor López-Monroy, Manuel Montes-y-Gómez, , Hugo Jair Escalante, , Luis Villaseñor-Pineda, Efstathios Stamatatos. Knowledge-based Systems, available online: July 2, 2015.



The article presents a detailed description and analysis of the methods developed by Pastor Lopez for winning the author profiling track of PAN 2013 and PAN 2014 challenges.



New paper about the baseline for the One-shot gesture recognition challenge

A new paper has been published online in the Journal of Pattern Analysis and Applications:

Principal motion components for one-shot gesture recognition. Hugo Jair Escalante, Isabelle Guyon, Vassilis Athitsos, Pat Jangyodsuk, Jun Wan. Pattern Analysis and Applications, available online: May 8, 2015.



This paper describes the baseline developed for the Chalearn one-shot gesture recognition challenge. The article presents an in depth description/analysis of the method which is based on motion histograms and PCA.



New paper on Term-weighting Learnin via Genetic programming

A new paper has been accepted for publication in the Journal of Knowledge-based Systems:

Term-Weighting Learning via Genetic Programming for Text Classification. Hugo Jair Escalante, Mauricio García-Limón, Alicia Morales-Reyes, Mario Graff, Manuel Montes-y-Gómez, Eduardo F. Morales, and José Martinez-Carranza. Knowledge-based Systems, available online: March 29, 2015.




New paper on Prototype Generation with MultiObjective Optimization

A new paper has been accepted for publication in the Pattern Analysis and Applications Journal:

MOPG: a multi-objective evolutionary algorithm for prototype generation. Hugo Jair Escalante,Maribel Marin-Castro,Alicia Morales-Reyes,Mario Graff,Alejandro Rosales-Pérez,Manuel Montes-y-Gómez,Carlos A. Reyes,Jesus A. Gonzalez Pattern Analysis and Applications, available online: February 5, 2015.




ChaLearn Board Meeting at NIPS2014

The board of directors of the Challenges in Machine Learning Organization (Chalearn) meet during the First Workshop on Challenges in Machine Learning, Montréal, Canada, December 12, 2014.



Journal paper accepted for publication in the Journal of Intelligent Manufacturing

An article reporting a solution based on machine learning to fault-bearing detection has been accepted for publication. The paper is a joint work with my colleagues Mario A. Saucedo now at Rochester Institute of Technology (formerly at PISIS-UANL) and Arturo Berrones from PISIS - UANL.

Detection of defective embedded bearings by sound analysis: a machine learning approach. Mario A. Saucedo-Espinosa, Hugo Jair Escalante, and Arturo Berrones. Journal of Intelligent Manufacturing, Published online: November 2014.



SNAIC - 2014 Call for participation to the 2do Seminario Nacional de Aprendizaje e Inteligencia Computacional - SNAIC 2014. SNAIC will be held at INAOE (Puebla-Mexico) on November 19-21, 2014.



Journal papers accepted for publication in Neurocomputing and IJPRAI

Three Journal papers on Evolutionary-based Machine Learning have been accepted for publication:

Multi-Objective Model Type Selection. A. Rosales-Pérez, J. A. Gonzalez, C. A. Coello Coello, H. J. Escalante, C. A. Reyes-Garcia. Neurocomputing, Available online: July 2011

Learning to Assemble Classifiers via Genetic Programming. N. Acosta-Mendoza, A. Morales-Reyes, H. J. Escalante, A. Gago-Alonso. International Journal of Pattern Recognition and Artificial Intelligence, Accepted, 2014

Surrogate-Assisted Multi-Objective Model Selection for Support Vector Machines. A. Rosales-Pérez, J. A. Gonzalez, C. A. Coello Coello, H. J. Escalante, C. A. Reyes-Garcia. Neurocomputing, Accepted 2014.

These papers report research lead by Alejandro Rosales and Niusvel Acosta.



Paper published in IEEE Transactions on Image Processing

A paper reporting work made by Jun Wan has been accepted for publication in the prestigious Journal IEEE Transactions on Image Processing:

J. Wan, V. Athitsos, P. Jangyodsuk, H.J. Escalante, Q. Ruan, I. Guyon. CSMMI: Class-Specific Maximization of Mutual Information for Action and Gesture Recognition. IEEE Transactions on Image Processing, Vol. 23(7) :3152--3165, July, 2014

The paper describes an approach for gesture and action recognition based on a vocabulary that is constructed by the maximization of intra and inter class mutual information. An extensive experimental evaluation is performed over several benchmark collections.



Papers accepted at GECCO and CEC 2014 Four papers have been accepted for publication in the top venues on evolutionary computation: GECCO and CEC. The accepted papers are:

  • Simultaneous Generation of Prototypes and Features through Genetic Programming. Mauricio García-Limón, Hugo Jair Escalante, Eduardo Morales, Alicia Morales. Accepted for GECCO 2014, full paper.
  • An Evolutionary Multi-Objective Approach for Prototype Generation. Alejandro Rosales, Hugo Jair Escalante, Jesus A. Gonzalez, Carlos A. Coello, Carlos A. Reyes. Accepted WCCI’2014, full paper.
  • Adaptive Surrogates with aNeuro-Fuzzy Network and Granular Computing. Israel Cruz Vega, Hugo Jair Escalante. Accepted for GECCO 2014, full paper.
  • Genetic programming of text representations. Mauricio García-Limón, Hugo Jair Escalante, Manuel Montes-y-Gómez, Alicia Morales, Eduardo Morales. Accepted for GECCO 2014, late breaking abstract.



New Journal papers accepted 2014 New research to be published in prestigious journals. This research is product of collaborations with colleagues from ChaLearn, Universidad de Michoacan and PISIS-UANL. The accepted papers are:



Award granted to Pastor López at CLEF-PAN'13 Pastor López-Monroy (PhD student directed by Manuel Montes-y-Gómez and myself), developed the method that obtained the best performance of the Author Profiling task at PAN'13. A total of 66 teams registered for the competition 21 of which submitted software for the final phase. Pastor's runs obtained the best average performance over English and Spanish corpora. Additionally, Pastor received a prize in memoriam of M. Teresa Turell sponsored by ForensicLab of the Universitat Pompeu Fabra, Barcelona. Pastor´s method implements second-order representations, which can capture semantic information and besides are highly efficient. The method is described in detail in the following publication:

Pastor López-Monroy, Manuel Montes-y-Gómez, Hugo Jair Escalante, Luis Villaseñor-Pineda, and Esaú Villatoro-Tello. INAOE’s participation at PAN’13: Author Profiling task. CLEF 2013 Evaluation Labs and Workshop – Working Notes Papers, 23-26 September, Valencia, Spain, 2013.



Contact:

Dr. Hugo Jair Escalante
Titular researcher A
Department of Computational Sciences
National Institute of Astrophysics, Optics and Electronics
Calle Luis Enrique Erro # 1,
Tonantzintla, Puebla, 72840, Mexico
Phone: +52 (222) 2663100 x 8319
Fax: +52 (222) 2663152
email: hugojair [at] inaoep [dot] mx
mirror website: http://hugojair.org