[Elsnet-list] Job Offer: PhD / Postdoc, Uni Magdeburg, Germany

Andreas Wendemuth andreas.wendemuth at e-technik.uni-magdeburg.de
Fri Jun 15 16:50:43 CEST 2007


Dear List,

the chair for cognitive systems (Prof. Andreas Wendemuth) at University
of magdeburg offers with immediate effect 4 positions for PhD/ Postdocs.
Salary is a researchers position on German public sector scale TVL 13.

Position 1: (for 6 years)
Position 2: Ph D student (for 6 years)
areas of speech recognition, speaker recognition, dialogue
systems, signal processing , cognitive modelling.
Position 3 (for 3 years)
"Multi-modal emotion recognition and blind source separation"
Position 4: (for 1 year)
"Context-dependent associative learning"


Details and contact see below:

Position 1: (for 6 years)
Junior Research Group Leader / Postdoc
covering the spectrum (1 or several of) of speech recognition, speaker 
recognition, dialogue
systems, signal processing , cognitive modelling.
See
http://iesk.et.uni-magdeburg.de/ko/fo.html
for a list of research topics

Position 2: Ph D student (for 6 years)
areas of speech recognition, speaker recognition, dialogue
systems, signal processing , cognitive modelling.
See
http://iesk.et.uni-magdeburg.de/ko/fo.html
for a list of research topics

Details for pos 1 and 2 can be found (in German) at
http://iesk.et.uni-magdeburg.de/ko/jobs/stelle-haushalt-magdeburg.html

Position 3 (for 3 years)
"Multi-modal emotion recognition and blind source separation"
Tasks:
1.Multimodality:
Extension of existing automatic speech recognition to combined 
acoustic/visual recognition under
poor conditions, definition of suitable confidence measures. 
Identification of suitable multi-modal
features for speech recognition with combined listening/lip reading. 
Optimization and dimensional
reduction by linear discrimination analysis.
2. Prosody:
The extraction of features from speech and video sequences result in a 
stream of prosodic and
facial feature units. Challenges arise, when analysing a sequence of 
speech and images, showing a
moving speaking person. The image data might not only describe the 
static appearance of the face
but also the dynamic behavior.
3. Classification:
Using prosodic and facial feature units as a teacher signal, we will 
train suitable classification
algorithms (e.g. neuronal network, support vector machines, hidden 
Markov models HMM) to
recognize emotional states in longer time periods. As an initial step, 
feature unit vectors can be
averaged over time windows.
4.Blind source separation:
Our working hypothesis is that emotional expressions in speech and face 
can be described in terms
of relatively few source variables, which can be defined by blind source 
separation (ICA). These
source variables may map on physiological states in the brain and 
autonomic nervous system.

Position 4: (for 1 year)
"Context-dependent associative learning"
Tasks:
. Is the learning of context-conditional associations by human observers 
influenced by, or even
predicated on, consistent temporal ordering of environmental events? In 
other words, can the
context-dependence of human associative learning be understood in terms 
of a temporalorder-
dependence?
. How does temporal-order-dependent learning compare to abstract 
learning algorithms (e.g.,
support-vector machines, dynamic adaptation of neural nets) for 
detecting patterns and
regularities in high-dimensional data streams?
. Is temporal-order-dependent learning suited as a general solution to 
complex learning
problems? How does it perform on diverse problems such as those 
described in section 7.3
(i.e., learning to recognize prosodic signals in speech or emotional 
markers in facial
expression)?

Positions 3 and 4 are within the
Bernstein-Group "Components of cognition: small networks to flexible rules".
http://neurosci.uni-magdeburg.de/bernstein/
Work period may be extended given continuous funding.

Details for pos 3 and 4 can be found (in German) at
http://iesk.et.uni-magdeburg.de/ko/jobs/stelle-bernstein.html

Contact:

Andreas Wendemuth

------------------
Prof. Dr. Andreas Wendemuth
Chair for Cognitive Systems / Speech Recognition
Otto-von Guericke University
FEIT IESK Building 2/3 Room 331
PO Box 4120
D- 39016 Magdeburg, Germany
phone: ++49 391 671844-8 / secr. -7
fax:   ++49 391 672005-2 / secr. -1
Andreas.Wendemuth at e-technik.uni-magdeburg.de
http://iesk.et.uni-magdeburg.de/ko/
location map:
http://www.uni-magdeburg.de/guests/ueb/icons/campus-uniplatz.gif
----------------------------------------------------





More information about the Elsnet-list mailing list