Breaker

Academic Studies & Courses

After, I spent some time to implement “Bresenham’s line” and “seed fill” algorithms for Image Processing course homework; I noticed that I haven’t blogged enough for my graduate program.

It is MS program in Computer Engineering at Istanbul University. I registered 3 courses; Advanced Computer Architectures, Image Processing Techniques for Computer Engineering, Computer Language Engineering for fall semester.  You can access course details by this link.

I haven’t decided my thesis subject yet but, I feel that it seems to be about Semantic Web, AI (like knowledge presentation, machine learning …) or Information Retrieval (IR) systems.

Nowadays, I’m dealing with Latent Semantic Analysis (LSA) or Latent Semantic Indexing (LSI).  It is a new subject but growing fast among next generation web search engines (like Google, Hakia … etc).  It needs some mathematical skills like linear algebra to analyze data. If I had been told this relation before I would payed more attention to mathematics than I did. Anyway, my mathematical knowledge is still fine 😉 I can find (solve) what I need (for now).  I’ll try to write much about my academic studies and interests.

merturk

2 thoughts on “Academic Studies & Courses

  • CJ

    Hi,

    you are wrong about LSI, it’s rather old, dating back to 1988, and it has quite important limitations. Newer systems are using probabilistic latent semantic indexing (PLSI). You’ll find that the “search engine optimisation” community have done an awful lot of analysis in this area regarding Google.

    Semantic web is more about ontologies, which is what Hakia use and cognition search also, to be basic about it. Google isn’t a semantic search engine, but they look for semantic relationships nonetheless, just that’s not what they rely on. It’s really very complicated, no search engine will rely on simply one algorithm. Google has over 200, of which PageRank.

    The Ubiquitous web is also part of web 3.0 is also very interesting, it goes hand in hand with semantic web.

    I don’t mean to be rude or anything, I’m finishing my PhD in natural language systems and built search engines for a year or so. I’m also a search engine optimisation expert.

    Check out these resources:

    http://www.seobook.com/archives/000657.shtml (2005)
    http://www.seobythesea.com/?p=1144 (loads of Google algo information)
    http://www.cognition.com/
    http://scienceforseo.blogspot.com/2008/11/why-plsi-is-prefered-over-lsi.html

    The last link is from my own blog, I write about IR papers all the time if you think it might be of interest.

    Good luck with your course and enjoy researching your thesis 🙂

    cj

  • Mehmet Ali Ertürk

    Hi CJ,

    Thanks for your reply. I’m quite new to these subjects (as you can guess) and looking for some help to eliminate my mistakes.

    I thought (or read somewhere) Google uses its own data structure to hold data like liked data structure (not RDF but like RDF). They may not use Semantic Web or Linked Data but they are good at constructing relations between websites.

    Hakia made integration with Yahoo! Search BOOS, to pay more attention to semantic researches, like SemanticRank algorithm (it’s my opinion). Actually, I don’t use Hakia as a search engine but steal good ideas.

    If you are finishing your PhD you are in the same position as my teachers, I take it as a suggestion, thanks for help. I already bookmarked the links, I will read bottom to top.

    Good luck for your studies 😉

Leave a Reply

Your email address will not be published. Required fields are marked *