**Seminar
on ****information theory and signal reconstruction **

__ Location:__
Max-Planck-Institute
for Astrophysics, Garching

__ Dates:
__Tuesday,
13. November & Friday, 23. November, 13:15--16:30; common lunch:
meeting 12:15 at MPA reception

**Agenda:**

Tuesday, 13. November:

1)
Javier Cuesta: **Quantum Information Theory**

2) Yi-Hao Chen:
**Neural Networks I**

3) Thomas Simon: **Fundamentals of Game
Theory**

4) Florian Lehner: **Poker from a Game Theoretical
Perspective**

Friday, 23. November:

1)
Philipp Seifert: **Kalman Filter**

2) Martin Spiessl: **Neural
Networks II**

3) Christian Huber: TBD

4) Felix ???: TBD

The
seminar is intended for participants of the lecture on *information
theory and signal reconstruction*,
the content of which will be assumed to be known by all participants.
The
main seminar goal is to extend the participants' knowledge beyond the
material covered in the lecture, especially with respect to concrete
measurement situations, imaging, and existing algorithms.

A second goal is to practice presentations and open discussions.

Talks
can be given at the blackboard, or by using computer projection and
should be in English and of 15 to 30 minutes length.

Topics should
be coordinated with the seminar leaders (mainly with Torsten Enßlin,
Niels Oppermann, Marco Selig).

A pdf-file with slides or write-up
of presented material (readable handwriting is ok) must be provided
by the student beforehand.

Computer presentations must be sent in
via email on the day before the seminar talk, so that they can be put
on a single presentation computer.

Two or three all-afternoon seminar dates will be fixed, probably one within the semester break, and the others during the WS 2012.

A marked certificate for 3 ECTS points will be provided to any student who

attended at least one of the seminar dates

gave an acceptable talk as specified here

and provided a pdf file with the talk slides or write-up to document her/his participation.

Topics can be picked from the list below, or chosen according to the students own interests. The talks should

make contact with the subject of the lecture

capture the essential of its topic rather than try to be exhaustive

extend the knowledge of the audience

do not have major overlap with the talks of last year's seminar on the same topic, which can be found here.

It is expected that the students prepare the talks on their own, without supervision. If difficulties with a topic arise during the preparation (e.g. suitable literature cannot be found, topic is too difficult to be presented or too trivial to be interesting) the student should change the topic of the talk to circumvent the problem. If she/he decides for a complete change of the topic, the seminar leaders should be informed via email. Literature research is part of the preparation work of the student. Teamwork and splitting of topics into several coordinated talks is welcome. Students are encouraged to have test talks with co-students beforehand.

Marks will be given according to the main criterion how informative the talk was given the lecture background. The quantity of presented material will be less important than the quality of the information transferred.

**Suggestions
of potential topics (variations of those and completely other topics
are also welcome):**

Medical imaging devices (e.g. description of measurement/ data model / likelihood)

Computer Tomography (CT)

Positron Emission Tomography (PET)

Nuclear Magnetic Resonance (NMR)

Functional Magnetic Resonance Imaging (fMRI)

Ultrasound imaging

...

Astronomical imaging devices (e.g. description of measurement/ data model / likelihood)

Optical telescopes

X-ray telescopes

Gamma-ray telescopes (coded mask imaging)

Neutrino telescopes

Ultra-high-energy arrays

Single-dish radio telescopes

Radio interferometer

CMB satellites (WMAP/Planck)

Geological probes (e.g. description of measurement/ data model / likelihood)

seismic waves

gravitational field measurements

Cartography via Kriging and other methods

...

Signal reconstruction / filter algorithms (e.g. algorithm / applications of it / implicit prior assumptions)

Richardson Lucy-Algorithm

Kalman Filter

Compressed Sensing

CLEAN Algorithm of Högbom (1947)

'

*Maximum Entropy Method'*of Imaging (Gull & Daniell, Nature 272, 686 (1978))Image compression (jpeg)

Audio compression (mp3)

Noise suppressing audio filter (dolby, …)

Specific signals (physics / statistical characteristics / prior models / information sources)

Cosmic Microwave Background radiation

Climate data

Stock market prices

1/f noise (e.g in transistors)

Applications of information theory (basic idea / simple cases)

Communication theory (Shannon, optimal coding, …)

Game theory (e.g. prisoner's dilemma, …) (talk by Thomas Simon)

Weather forecasts (data assimilation)

Parameter estimation (e.g. in cosmology)

Loan granting decisions of banks

Insurance mathematics

Quantum information theory (talk by Javier Cuesta)

Language processing