Skip to main content

Semi-supervised Structured Output Learning from Partially Labeled Data

Objective

Learning classifiers automatically from examples is subject to the
multidisciplinary field of machine learning.

The structured output learning (SOL) is concerned with the
learning of classifiers for prediction of multiple
interdependent variables exhibiting some structure dependence.
Recent progress in SOL focuses mainly on supervised methods
that require labeled examples. A high cost of labeled examples
significantly limits application of SOL to many domains.

Our goal is threefold. First, to developed framework for semi-supervised SOL from cheap partially labeled examples. Second, to apply this new framework to two important SOL tasks: (i) Markov Networks learning and (ii) learning of 2-dimensional image grammars. Third, to use the new algorithms for solving computer vision problems including the image segmentation and the car license plate recognition.

To achieve the first goal, we will examine two strategies. First, we will
combine powerful discriminative methods for SOL with generative models offering a principled way to deal with missing labels. Second, we will extend the existing semi-supervised methods in order to handle the partially labeled examples.

To achieve the second goal, we will incorporate the existing methods for
supervised SOL of Markov Networks and 2D grammars to the framework
developed as the first goal.

To achieve the third goal, we will build on the technology for
image segmentation and license plate recognition developed by
the host. The currently used classification methods will be
replaced by the developed semi-supervised SOL algorithms to
demonstrate their effectiveness on real-life problems.

Achieving the goals will be possible by joining the expertise
of the applicant and the host. This applies both to theoretical
and application oriented goals. The applicant is experienced in
SOL and Markov Networks while the host will complement this
with a large expertise in 2D grammars and computer
vision.

Field of science

  • /natural sciences/computer and information sciences/artificial intelligence/computer vision
  • /natural sciences/computer and information sciences/artificial intelligence/machine learning

Call for proposal

FP7-PEOPLE-ERG-2008
See other projects for this call

Funding Scheme

MC-ERG - European Re-integration Grants (ERG)

Coordinator

CESKE VYSOKE UCENI TECHNICKE V PRAZE
Address
Jugoslavskych Partyzanu 1580/3
160 00 Praha
Czechia
Activity type
Higher or Secondary Education Establishments
EU contribution
€ 45 000
Administrative Contact
Igor Mraz (Mr.)