Kelwin Fernandes

Phd Student at University of Porto & INESC TEC

Name Kelwin Fernandes
Date of birth 1990/02/22
Address Porto, Portugal
Languages Spanish, English, Portuguese
Email kafc at inesctec dot pt

Kelwin Fernandes My name is Kelwin Fernandes and I am a PhD student in Informatics at University of Porto and at INESC TEC in Porto, Portugal.

I obtained my Bsc degree in Computer Engineering with Summa Cum Laude distinction at Universidad Simón Bolívar in Caracas, Venezuela.

My main research insterest include: Machine Learning, Computer Vision and Artificial Intelligence.

Download PDF resume

2014 - 2018 (expected)

Universidade do Porto

PhD Student in Informatics

  • Thesis title: Cervical Cancer Detection using Digital Colposcopies
  • Supervisor: Jaime S. Cardoso
  • GPA: 18.75/20

2007 - 2012

Universidad Simón Bolívar

Computer Engineering (Summa Cum Laude)

  • Thesis title: Teeth and palate detection in palatal view intraoral images
  • Supervisor: Carolina Chang
  • GPA: 4.84/5 (Class rank: 1st/248)

Learning to Rank
Abstract: We study the problem of learning lexicographic preferences on multiattribute domains, and propose Rankdom Forests as a compact way to express preferences in learning to rank scenarios. We start generalizing Conditional Lexicographic Preference Trees by introducing multiple kernels in order to handle non-categorical attributes. Then, we define a learning strategy for inferring lexicographic rankers from partial pairwise comparisons between options. Finally, a Lexicographic Ensemble is introduced to handle multiple weak partial rankers, being Rankdom Forests one of these ensembles. We tested the performance of the proposed method using several datasets and obtained competitive results when compared with other lexicographic rankers. (code, data, results and paper)

Imbalanced Data
Abstract In classification, when there is a disproportion in the number of observations in each class, the data is said to be class imbalance. Class imbalance is pervasive in real world applications of data classification and has been the focus of much research. The minority class contributes too little to the decision boundary because the learning process learns from each observation in isolation. We discuss the application of learning pairwise rankers as a solution to class imbalance. We compare ranking models to alternatives from the literature. (code, data, results and paper)

Directional Data
Abstract: In different areas of knowledge, phenomena are represented by directional -angular or periodic- data; from wind direction and geographical coordinates to time references like days of the week or months of the calendar. These values are usually represented in a linear scale, and restricted to a given range (e.g. [0, 2π)), hiding the real nature of this information. Therefore, dealing with directional data requires special methods. So far, the design of classifiers for periodic variables adopts a generative approach based on the usage of the von Mises distribution or variants. Since for nonperiodic variables state of the art approaches are based on nongenerative methods, it is pertinent to investigate the suitability of other approaches for periodic variables. We propose a discriminative Directional Logistic Regression model able to deal with angular data, which does not make any assumption on the data distribution. Also, we study the expressiveness of this model for any number of features. Finally, we validate our model against the previously proposed directional naïve Bayes approach and against a Support Vector Machine with a directional Radial Basis Function kernel with synthetic and real data obtaining competitive results.

  1. Discriminative Directional Classifiers (Kelwin Fernandes, Jaime S. Cardoso), In Neurocomputing, 2016.

  2. Learning and Ensembling Lexicographic Preference Trees with Multiple Kernels (Kelwin Fernandes, Jaime S. Cardoso, Hector Palacios), In Proceedings of International Joint Conference on Neural Networks (IJCNN), 2016.

  3. Tackling Class Imbalance with Ranking (Ricardo Cruz, Kelwin Fernandes, Jaime S. Cardoso, Joaquim F. Pinto da Costa), In Proceedings of International Joint Conference on Neural Networks (IJCNN), 2016.

  1. Temporal Segmentation of Digital Colposcopies (Kelwin Fernandes, Jaime S. Cardoso, Jessica Fernandes), In Proceedings of Iberian Conference on Pattern Recognition and Image Analysis (IbPRIA), 2015.

  2. A Proactive Intelligent Decision Support System for Predicting the Popularity of Online News (Kelwin Fernandes, Pedro Vinagre, Paulo Cortez), In Progress in Artificial Intelligence(IbPRIA), 2015 (Best Paper Award).

  1. Pavement pathologies classification using graph-based features (Kelwin Fernandes, Lucian Ciobanu), ICIP, 2014.

  2. Catalogue-Based Traffic Sign Asset Management: Towards User’s Effort Minimisation (Kelwin Fernandes, Pedro Silva, Lucian Ciobanu, P. Fonseca), In Proceedings of International Conference on Image Analysis and Recognition (ICIAR), 2014.

  1. A* mbush Family: A* Variations for Ambush Behavior and Path Diversity Generation (Kelwin Fernandes, Glebys González, Carolina Chang), In Proceedings of Motion in Games (MIG), 2012.

  2. Teeth/Palate and interdental segmentation using artificial neural networks (Kelwin Fernandes, Carolina Chang), Artificial Neural Networks in Pattern Recognition (ANNPR), 2012.
  1. Best Paper Award at EPIA 2015, Coimbra, Portugal.

  1. CTM Researcher-Entrepreneur Competition, INESC TEC. Portugal.

  1. PhD Grant (SFRH/BD/93012/2013), FCT, Portugal.
    Third best candidature in Computer Science in Portugal

  1. Outstanding Performance Award as a Teaching Assistant, Universidad Simón Bolívar.

  2. Best academic average of cohort 2007 in Computer Engineering, Universidad Simón Bolívar.

  1. Best Student Award, Universidad Simón Bolívar.

  2. Best academic average of cohort 2007 in Computer Engineering, Universidad Simón Bolívar.

  1. Honor of Merit 2007, Universidad Simón Bolívar.
Universidade do Porto
  1. Lab. of Programming 2 (Data Structures)

Universidad Simón Bolívar
  1. Mathematics I (Differential Calculus)
  2. Symbolib Logic - three periods
  3. Lab. of Algorithms and Data Structures I (Introduction to Programming)
  4. Lab. of Algorithms and Data Structures II (Data Structures)
  5. Lab. of Algorithms and Data Structures III (Graph Theory)
  6. Artificial Intelligence II (Machine Learning)