Post-doc in machine learning for semantic composition at the University of Trento

Post-doc in machine learning for semantic composition at the University of Trento

Posted by Victoria Nicholl on Tue, 24/01/2012 - 10:10

One (RENEWABLE) 2-YEAR POST-DOC POSITION IN MACHINE LEARNING AVAILABLE

The CIMeC-CLIC laboratory of the University of Trento, an
interdisciplinary group of researchers studying language and
conceptualization using both computational and cognitive methods
(clic.cimec.unitn.it) announces the availability of a 2-year Post-Doc
position in machine learning, renewable up to a maximum of 4 years.

The scholarship is funded by a 5-year European Research Council
Starting Grant awarded to the COMPOSES (COMPositional Operations in
SEmantic SPACE) project (clic.cimec.unitn.it/composes), that aims to
model the meaning of phrases and sentences with computational methods.

* Research Goals and Desired Profile *

Distributional semantics is a general framework to induce vector-based
meaning representations of words from collections of naturally
occurring text (corpora) on a large scale. The successful candidate
will develop, in collaboration with the COMPOSES project team, novel
machine learning techniques to derive distributional semantic
representations of phrases and sentences from distributional
representations of words and other corpus data (e.g., deriving "red
dog" from corpus-based representations of "red" and "dog"). To achieve
this goal, we face the hard challenge to learn output representations
that are very high-dimensional vectors from inputs that are also
high-dimensional vectors, that might in turn be the output of other
empirically-learned functions.

The successful candidate should have experience in one or more of the
following areas: regularization methods, hierarchical regression,
dimensionality reduction and/or feature selection for multidimensional
multiple regression learning, scaling machine learning to large
multivariate and multi-level problems, dealing with very sparse data,
efficient large-scale implementation of regression methods, learning
algorithms for deep architectures. The research fellow must also have
a strong interest in working in an interdisciplinary environment.

* The Research Environment *

The CLIC lab (clic.cimec.unitn.it) is a unit of the University of
Trento's Center for Mind/Brain Sciences (CIMeC,
www.unitn.it/en/cimec), an English-speaking, interdisciplinary center
for research on brain and cognition whose staff includes
neuroscientists, psychologists, (computational) linguists, computer
scientists and physicists.

CLIC consists of researchers from the Departments of Computer Science
(DISI) and Cognitive Science (DISCoF) carrying out research on a range
of topics including concept acquisition, corpus-based computational
semantics, combining NLP and computer vision, combining brain and
corpus data to study cognition, formal semantics and theoretical
linguistics. Modeling composition in distributional semantics is
increasingly a focus point of CLIC, and activity in this area will
grow considerably thanks to COMPOSES funds.

CLIC is part of the larger network of research labs focusing on
Natural Language Processing and related domains in the Trento region,
that is quickly becoming one of the areas with the highest
concentration of researchers in NLP and related fields anywhere in
Europe.

The CLIC/CIMeC laboratories are located in beautiful Rovereto, a
lively town in the middle of the Alps, famous for its contemporary art
museum, the quality of its wine, and the range of outdoors sport and
relax opportunities it offers:

http://en.wikipedia.org/wiki/Rovereto

* Application Information *

For further information, please send an expression of interest to
marco.baroni@unitn.it, attaching a CV. The position is available
immediately and open until filled.