Semi-supervised Learning via Generalized MaxEntMay 2008 - December 2009
From May 2008 until December 2009, I was a research fellow at the Max Planck Institute for Biological Cybernetics, Department Shoelkopf. As a visiting PhD student at MPI, I worked on semi-supervised algorithms for structured output prediction with Dr. Yasemin Altun.
Advised by Dr. Yasemin Altun, Max Planck Institute for Biological Cybernetics, Department Shoelkopf, Tuebingen, Germany.
A Maximum Entropy Approach to Semi-Supervised Learning
Ayse Erkan, Yasemin Altun, 30th International Workshop on Bayesian Inference and Maximum Entropy Methods in Science and Engineering, Chamonix, France, July 4-9, 2010
Semi-supervised Remote Sensing Image Classification via Maximum Entropy
Ayse Erkan, Gustavo Cammps-Valls, Yasemin Altun, MLSP Workshop, 2010
Semi-supervised Learning via Generalized Maximum Entropy
Ayse Erkan, Yasemin Altun, AI & Statistics (AISTATS), 2010
Learning Probabilistic Models of Grasp AffordancesJuly 2009 - September 2009
In this project I worked on the problem of learning and efficiently representing discriminative probabilistic models of object-specific grasp affordances with minimal labeled grasp configurations. We train Kernel logistic regression (KLR) to map the hypothesis space of grasps into continuous class conditional probability values indicating their achievability.
We define a distance metric and an associated
kernel that combine 3D position and orientation
features in the object relative reference
frame. While the hypothetical configurations
acquired with the 3D visual model are abundant,
labeled configurations are very limited as these
are obtained via time-costly experiments
controlled by a human observer. We propose a
semi-supervised extension of KLR and a framework
to combine the merits of semi-supervised and
active learning approaches to tackle the scarcity
of labeled grasps. Experimental evaluation shows
that combining active and semi-supervised
learning is favorable if there is access to an
oracle, semi-supervised learning outperforms
supervised learning otherwise, particularly when
the labeled data is very limited.
Joint work with Oliver Kroemer, Renaud Detry, Jan Peters, Justus Piater, and Yasemin Altun.
Learning Probabilistic Discriminative Models of Grasp Affordances under Limited Supervision
Ayse Erkan, Oliver Kroemer, Renaud Detry, Yasemin Altun, Justus Piater, Jan Peters, International Conference on Intelligent Robots and Systems (IROS), 2010
Large Scale Manifold TransductionMay-August 2007
During summer 2007, as a research intern I worked on online large scale transduction methods at NEC Labs, Princeton.
Advised by Dr. Jason Weston, Dr. Ronan Collobert, NEC Labs Princeton, NJ
Large Scale Manifold Transduction
Michael Karlen, Jason Weston, Ayse Erkan, and Ronan Collobert, International Conference on Machine Learning (ICML) 2008
Learning Applied To Ground Robots (LAGR)December 2004 – January 2008 (Project page @ CBLL)
is a project funded by the Defense Advanced Research Projects
Agency (DARPA), which aims to promote the development of better
learning systems for robot navigation in unconstrained outdoor
environments. Several teams compete to get the best improvement to the
baseline system developed by Carnegie Mellon University. The goal is to
reach a given destination in the shortest time, using primarily two
pairs of cameras as sensors, which makes obstacle avoidance in the
robot’s course the main challenge.
Our lab, in collaboration with
Net-Scale technologies, Inc., developed a long range
obstacle detection system that allows the robot to recognize obstacles
at up to 35m. In this project, I implemented a self-supervised label
propagation scheme that uses location correspondences and thus enables
learning by using features gathered from various views of the same
obstacle, i.e., from different orientations, scales, and lighting
Advised by Prof. Yann LeCun Computational and Biological Learning Lab (CBLL), New York University.
No Directions Required--Software
Smartens Mobile Robots
Peter Sergo. Scientific American, by Feb 21st, 2008.
Learning long-range vision for autonomous off-road driving
Raia Hadsell, Pierre Sermanet, Jan Ben, Ayse Erkan, Marco Scoffier, Koray Kavukcuoglu, Urs Muller, Yann LeCun Journal of Field Robotics
Volume 26 Issue 2 , Pages 120 - 144 (February 2009)
Deep Belief Net Learning in a Long-Range Vision System
for Autonomous Off-Road Driving
Raia Hadsell, Ayse Erkan, Pierre Sermanet, Marco Scoffier, Urs Muller, Yann LeCun. Intelligent Robots and Systems (IROS) 2008
Self-Supervised Learning From High Dimensional Data for Autonomous Off-Road Driving
Ayse Erkan, Raia Hadsell, Pierre Sermanet, Koray Kavukcuoglu, Marc'Aurelio Ranzato, Urs Muller, Yann LeCun. Presented at NIPS 2007 Workshop: Robotic Challenges for Machine Learning
Adaptive Long Range Vision in Unstructured Terrain
A. Erkan, R. Hadsell, P. Sermanet, J. Ben, U. Muller, Y. LeCun, Intelligent Robots and Systems (IROS) 2007
A Multi Range Vision Strategy for Autonomous
Off Road Navigation
R. Hadsell, A. Erkan, P. Sermanet, J. Ben, K. Kavukcuoglu, U. Muller, Y. LeCun, IASTED International Conference on Robotics and Applications (RA) 2007.
Speed Range Dilemmas for Vision Based
Navigation in Unstructured Terrain
P. Sermanet, R. Hadsell, J. Ben, A. Erkan, B. Flepp, U. Muller, Y. LeCun. Intelligent Autonomous Vehicles (IAV) 2007.
Online Learning for Off Road Robots: Using Spatial Label
Propagation to Learn Long Range Traversability
R. Hadsell, P. Sermanet, J. Ben, A. Erkan, J. Han, B. Flepp, U. Muller, Y. LeCun, Robotics Science and Systems (RSS) 2007.
Human Gesture RecognitionDecember 2002 – August 2004
Bogazici University Computer Science Department is a participant in the SIMILAR project (the European taskforce creating human-machine interfaces SIMILAR to human-human communication). The main objective of our group was to develop a human-computer interface for the disabled. I took part in the implementation of an online 3D Hand Gesture Recognition System using Hidden Markov Models, where the user controls PC applications with a colored glove. Later on, as part of my Master’s thesis, I worked on 3D Hand Posture Recognition using nonparametric optimization techniques and developed an application that distinguishes different postures of the hand such as those in sign language.
Advised by Prof. Lale Akarun, Perceptual Intelligence (PI) Lab, Bogazici University, Istanbul, TURKEY.
Based Three Dimensional Hand Posture Recognition for Hand Tracking
Ayse Naz Erkan, Master's Thesis, Bogazici University, 2004
Hand Tracking and 3D Gesture Recognition for Interactive Interfaces
C. Keskin, A. Erkan, L. Akarun, ICANN/ICONIP, Istanbul, June 2003
Arayuzler icin Gercek Zamanli El Izleme Ve HMM Tabanli Uc Boyutlu Hareket Tanima
A. Erkan, C.Keskin, L. Akarun, SIU, Istanbul, 2003 (in Turkish)
Cerberus: RoboCup 2002 Sony Legged Robot League ParticipantSeptember 2001 – June 2002
RoboCup is an
international research and education initiative that
fosters robotics research. Our
was a participant in the Sony Legged Robot League held
in Fukuoka, Japan in 2002. The project consisted of five
modules: vision, locomotion, planning and behavior,
localization, and communication. As part of my
undergraduate thesis, I took part in the localization
module where I implemented triangulation techniques
using the landmarks in the soccer field. I also
developed basic machine learning tools for object
detection as part of the vision module.
Advised by Prof. Levent Akin, Bogazici University, Istanbul, TURKEY.