Institut f�r Grundlagen der Informationsverarbeitung
(708)

**Leaders of the Seminar:**

Assoc. Prof. Dr. DI Robert Legenstein

O.Univ.-Prof. Dr. Wolfgang Maass

**Office hours:** by appointment (via e-mail)

E-mail:

maass@igi.tugraz.at

robert.legenstein@igi.tugraz.at

Homepage: https://igi-web.tugraz.at/people/maass/

Location: IGI-seminar room, Inffeldgasse 16b/I, 8010 Graz

time: Monday 4.15 pm

http://www.humanbrainproject.eu/

https://igi-web.tugraz.at/people/maass/nonexperts.html

in particular into the research area "Principles of Brain Computation" in this project, for which our Institute is responsible.

This seminar is designed for master students in their second year, or also in the first year, if they take simultaneously NN B (or have already

taken one of the courses listed below).

The papers and talks will be on an introductory level, and will be accessible to an audience with minimal technical background. One didactical goal

of the seminar is to provide experience in giving presentation, which can be practiced here in a relaxed setting.

In addition, students can explore research topics in which they might be interested for a master thesis or -project.

One of the courses NN A, NN B, ML A, ML B suffices as background.

10 .06. 2013 Florian Hofer, slides

10 .06. 2013 Wolfgang Roth, slides

Papers related to Principles of Brain Computation/Human Brain Projec

1.
Eliasmith, C., Stewart, T. C., Choo, X., Bekolay, T.,
DeWolf, T., Tang, C., & Rasmussen, D. (2012). A large-scale
model of the functioning brain. Science, 338(6111),
1202-1205.

http://clm.utexas.edu/compjclub/papers/Eliasmith2012.pdf

**Supplementary
Material**
PDF

**Commentary:**
Machens, C. K. (2012). Building
the Human Brain. Science, 338(6111), 1156-1157.
PDF

Three presentations should cover this
material.

2.
A model for stochastic computation in cortical
microcircuits (new
paper by Habenschuss,
Jonke, Maass),
preprint will be made available to seminar
participants

This paper should be covered by 2 talks (one on
stationary distributions of network states, one on Sudoku
application).

3.
Larkum, M. (2012). A cellular mechanism for cortical
associations: an organizing principle for the cerebral cortex.
Trends in Neurosciences. PDF

One talk can cover this paper.

4.
Sporns, Olaf. "Network attributes for segregation and
integration in the human brain." Current opinion in neurobiology
(2013). PDF

One talk can cover this paper.

5.
S. Haeusler and W. Maass. A statistical analysis of
information processing properties of lamina-specific cortical
microcircuit models. Cerebral Cortex, 17(1):149-162,2007.

https://igi-web.tugraz.at/people/maass/psfiles/162.pdf

One or two talks can be
given on this paper. (Possibly the talks can cover also some
related new results for a column model of the Human Brain
Project)

**Papers related to machine
learning and artificial neural
networks:**

1. Neal, R. (1992). Connectionist
learning of belief networks. Artificial Intelligence, 56:
71-113

http://www.sciencedirect.com/science/article/pii/0004370292900656

This paper introduces sigmoidal belief networks and
relates them to Boltzmann machines. This talk should be the basis
for a later talk on deep belief networks. (The PDF is not freely
available. It will be provided to seminar
participants)

2. Hinton, G. E., Osindero, S.
and Teh, Y. (2006). A fast learning algorithm for deep belief
nets. Neural Computation 18, pp 1527-1554.
2006.

http://www.cs.toronto.edu/%7Ehinton/absps/ncfast.pdf

Introduces greedy-layerwise training in deep belief
networks.

3.
Salakhutdinov R.
(2010). Learning Deep Boltzmann Machines using Adaptive MCMC.
Proc. of the 27^{th} International Conference on Machine
Learning.

http://www.cs.utoronto.ca/~rsalakhu/papers/adapt.pdf

Introduces a novel way to
learn deep Boltzmann machines (as opposed to DBNs).

4.
Hinton, G. E.,
Srivastava, N., Krizhevsky, A., Sutskever, I. and Salakhutdinov, R. (2012).
Improving neural networks by preventing co-adaptation of feature
detectors. Technical Report.

http://arxiv.org/abs/1207.0580

Introduces a novel
technique to prevent overfitting in deep feed-forward neural
networks.

5.
G. M. Hoerzer, R.
Legenstein, and Wolfgang Maass (2012). Emergence of complex
computational structures from chaotic neural networks through
reward-modulated Hebbian learning. Cerebral
Cortex.

https://igi-web.tugraz.at/people/maass/psfiles/214_incl_suppl.pdf

This paper is related to Echo State Networks that were
discussed briefly in NNA, but it is extended and discussed in a
biological context.

**Papers related to machine
learning and robotics research for the EU-project
AMARSi:**

1. Ijspeert,
A.;Nakanishi, J.; Pastor, P; Hoffmann, H.; Schaal, S. (2013).
Dynamical Movement

Primitives: Learning
Attractor Models for Motor Behaviors, *Neural Computation*, **25**, pp.328�373.

http://www-clmc.usc.edu/publications/I/ijspeert-NC2013.pdf

This journal paper provides a detailed
discussion on parametrized elementary movements using dynamical
systems. This approach is the most widely used movement primitive
representation in robotics.

2. Muelling, K.;
Kober, J.; Kroemer, O.; Peters, J. (2013). Learning to Select and
Generalize Striking

Movements in
Robot Table Tennis, *International Journal of Robotics
Research*

http://www.ias.informatik.tu-darmstadt.de/uploads/Publications/Muelling_IJRR_2013.pdf

This paper presents
a motor skill learning approach implementing a movement primitive
library. In impressing exeriments the authors show how a robot
learns to play table tennis.

3. Matthew Botvinick,
Marc Toussaint (2012):
Planning as Inference. *Trends in Cognitive
Sciences*,

16(10),
485-488, 2012.

http://ipvs.informatik.uni-stuttgart.de/mlr/marc/publications/12-BotvinickToussaint-TICS.pdf

Paper on how the human brain may implement
motor planning. The authors discuss recent advances in motor
planning using probabilistic models.

Can be combined with a more theoretical
machine learning paper, i.e. Konrad Rawlik, Marc Toussaint, Sethu
Vijayakumar: On stochastic optimal control and reinforcement
learning by approximate inference. *Int. Conf. on Robotics Science and
Systems* (R:SS 2012). *Best paper
runner up award.*
http://ipvs.informatik.uni-stuttgart.de/mlr/marc/publications/12-rawlik-toussaint-vijayakumar-RSS.pdf

4. Quadrupet Robots
and Motor Control Approaches: The following three papers should
be

combined. The
discuss robot modelling appraoches and trajectory formation
strategies:

Sproewitz A, Kuechler L, Tuleu A, Ajallooeian M, D�Haene M,
Moeckel R, Ijspeert AJ. (2011): Oncilla Robot, A Light-weight
Bio-inspired Quadruped Robot for Fast Locomotion in Rough
Terrain. In: Symposium on Adaptive Motion of Animals and Machines
(AMAM2011). Symposium on Adaptive Motion of Animals and Machines
(AMAM2011). ; 2011. p. 63-64.

http://infoscience.epfl.ch/record/182313/files/s323.pdf

A. Sproewitz, M. Fremerey, K. Karakasiliotis, S. Rutishauser and
L. Righetti (2009). Compliant Leg Design for a Quadruped Robot.
Dynamic Walking 2009, Vancouver, Canada.

http://infoscience.epfl.ch/record/142736/files/Sproewitz08_dw_preprint.pdf

S. Rutishauser, A. Sproewitz, L. Righetti and A. J. Ijspeert
(2008). Passive compliant quadruped robot using central pattern
generators for locomotion control. International Conference on
Biomedical Robotics and Biomechatronics, Scottsdale,
2008.

http://infoscience.epfl.ch/record/130727/files/sRutishauser08.pdf

2013/rh