Results 1 to 12 of 12

Thread: PhD studies

  1. #1
    Member
    Join Date
    Dec 2011
    Location
    Germany / Hessen
    Posts
    18
    Thanks
    0
    Thanked 1 Time in 1 Post

    PhD studies

    Hello everybody,

    I got a question to you:

    I just finished my master`s degree and now I am looking for a place where I could do my PhD.

    Did any of you do his PhD studies / thesis in the area of data compression?
    Or has somebody hints / links for scientific working groups or professors who are into this topic?
    (perhaps somebody has some interesting links / hints I didn`t find yet)

    Perhaps you can share some of your experience with this topic (PhD related to data compression).


    I would be very happy to hear from you. Any comments are welcome!

    Best regards,

    Steve

  2. #2
    Expert
    Matt Mahoney's Avatar
    Join Date
    May 2008
    Location
    Melbourne, Florida, USA
    Posts
    3,255
    Thanks
    306
    Thanked 778 Times in 485 Posts
    I wanted to do my Ph.D on text compression as a means to estimate the cost of AI. My thesis adviser had funding for something else, so after about a year of trying to work on the topic I was interested in and doing paid work on the other topic (intrusion detection), I realized I did not have time to do both so I changed my topic in order to graduate. After graduation I spent a few years developing PAQ and LTCB as a hobby while teaching as a part time adjunct, and eventually I was recruited for a better paying job in the field based on my open source work.

    Data compression is a highly specialized field. Probably only 100 people worldwide can write good compression algorithms and half of them are probably here. I couldn't tell you who is working on compression and teaching at a university.

  3. #3
    Member m^2's Avatar
    Join Date
    Sep 2008
    Location
    Ślůnsk, PL
    Posts
    1,612
    Thanks
    30
    Thanked 65 Times in 47 Posts
    Silesian University of Technology has been working on compression short time ago, they may be still doing it. I have to say that skimming through 1 thesis and 1 paper I haven't considered their work to be very good, though I want to be clear: it's just a first impression.

  4. #4
    Programmer toffer's Avatar
    Join Date
    May 2008
    Location
    Erfurt, Germany
    Posts
    587
    Thanks
    0
    Thanked 0 Times in 0 Posts
    Hi,

    i do research (and my PhD) in data compression and i'm working in this area for roughly 3 years.
    My research mostly is about optimization in statistical data compression, i.e. analyzing prediction models and mixture methods (including theoretical results on PAQ).
    Despite me i unfortunately only know about a single academic person in Germany, who works at a university and did research in data compression.
    But there're definitely some in other countries.

    Some hints:
    - Convince some professor and write an essay on your research plans.
    (This worked for me, but i was in the right place at the right time.)
    - Do different research, where you can use data compression methods.
    (E.g. sequential prediction).
    - I found that a lot of data compression problems can be cast into machine learning / online optimization problems.
    You may want to search for machine learning professors and do your PhD in this area, specializing into the data compression niche.
    - Do your PhD in a different country.

    Cheers
    Toffer
    M1, CMM and other resources - http://sites.google.com/site/toffer86/ or toffer.tk

  5. #5
    Expert
    Matt Mahoney's Avatar
    Join Date
    May 2008
    Location
    Melbourne, Florida, USA
    Posts
    3,255
    Thanks
    306
    Thanked 778 Times in 485 Posts
    Good point. I did some postdoc work in time series anomaly detection (predicting failures in valves used on the space shuttle). I showed that C(training)+C(test)-C(training+test) where C() is compressed size is smaller for anomalies (less mutual information between training and test data). Also the test is more reliable for better compressors like paq6 vs. gzip. But we ended up using other machine learning techniques for performance reasons.

  6. #6
    Member
    Join Date
    Oct 2010
    Location
    Germany
    Posts
    275
    Thanks
    6
    Thanked 23 Times in 16 Posts
    I did a small university project in the field of data compression. It was mainly a (theoretical) description of LZ77, PPM, PAQ etc, and it showed results of my own work.

    I am currently doing my PhD (inverse modelling) and have to admit, that, as I learn more and more about (stochastic) optimization, regression and machine learning I could do much better in compression

    So (text) compression is, as toffer already stated, an applied side product of other techniques, because in applied sciences it's all about prediction at the end
    Last edited by Sebastian; 22nd January 2014 at 10:49.

  7. #7
    Programmer Bulat Ziganshin's Avatar
    Join Date
    Mar 2007
    Location
    Uzbekistan
    Posts
    4,497
    Thanks
    733
    Thanked 659 Times in 354 Posts

  8. #8
    Member
    Join Date
    Oct 2010
    Location
    New York
    Posts
    21
    Thanks
    9
    Thanked 2 Times in 1 Post
    You might consider doing your PhD in Compressive (or Compressed) Sensing, a related and extremely active field of research. There is a substantial amount of funding available, many groups doing excellent research and a set of problems that overlap with data compression. Whereas generalized data compression is highly specialized, compressive sensing touches on multiple experimental fields.

  9. #9
    Member
    Join Date
    Dec 2011
    Location
    Germany / Hessen
    Posts
    18
    Thanks
    0
    Thanked 1 Time in 1 Post
    THX to everybody who has answered so far : )
    Your posts are very interesting for me.

    @toffer: are you doing your PhD in germany?

    @m^2 at least I found a PhD thesis at Silesian University related to data compression of 2003 "Deorowicz, S., Universal lossless data compression algorithms"
    but...well I am not quite sure, if a PhD in Poland is similar to one in Germany/US/UK - I will google this

    @Sebastian and Seth
    The idea of doing a PhD in related fields sounds good - this makes it a lot of easier to find a PhD position
    But
    perhaps also leads to the problem Matt had between funding and interests...I guess

  10. #10
    Programmer toffer's Avatar
    Join Date
    May 2008
    Location
    Erfurt, Germany
    Posts
    587
    Thanks
    0
    Thanked 0 Times in 0 Posts
    Yes, i'm doing my PhD in Germany, more specifically: tu-ilmenau.de.
    M1, CMM and other resources - http://sites.google.com/site/toffer86/ or toffer.tk

  11. #11
    Member
    Join Date
    Feb 2013
    Location
    San Diego
    Posts
    1,057
    Thanks
    54
    Thanked 71 Times in 55 Posts
    Quote Originally Posted by Matt Mahoney View Post
    I couldn't tell you who is working on compression and teaching at a university.
    Wouldn't you agree that the program committee for the DCC would give you a good start on making a list?
    http://www.cs.brandeis.edu/~dcc/index.html

    The list of presenters, also. Research is highly social, so if you could get in contact with one such academic, they could most likely point you toward others, and even give you inside tips on who is good.
    Last edited by nburns; 23rd January 2014 at 19:31.

  12. #12
    Member m^2's Avatar
    Join Date
    Sep 2008
    Location
    Ślůnsk, PL
    Posts
    1,612
    Thanks
    30
    Thanked 65 Times in 47 Posts
    Quote Originally Posted by Steve View Post
    @m^2 at least I found a PhD thesis at Silesian University related to data compression of 2003 "Deorowicz, S., Universal lossless data compression algorithms"
    but...well I am not quite sure, if a PhD in Poland is similar to one in Germany/US/UK - I will google this
    There has been more recent work, but I don't know precisely. You may want to mail Deorowicz to ask what's going on around.
    AFAIK there are no big differences between PhD conducts in EU, but sure, do your research.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •