ogrisel

26 ms·

1. 241.▲
   by ogrisel 15y ago

   I really enjoyed this night, and especially the vast array of topics we got to discuss either in "stand up" mode or in small groups with beers. A unique opportunity to keep up with the bleeding edge research done in Bayesian Data Fusion app
2. 242.▲
   A Berlin Buzzwords 2011 wrap up focusing on Hadoop, Machine Learning and NLP(blogs.nuxeo.com)

   4 points by ogrisel 15y ago|0 comments
3. 243.▲
   by ogrisel 15y ago

   Thanks for the feedback :)
4. 244.▲
   by ogrisel 16y ago

   You should cheer up that we are living in the Slow Zone. That should make us unaffected by the "Straumli Perversion" and other Powers from the Transcend and messing with the Beyond for quite some time.
5. 245.▲
   by ogrisel 16y ago

   Have you scrolled down the page a bit? There are commented source code samples and links to the generated outputs, source code and instructions to run it yourself if you wish.
6. 246.▲
   Mining Wikipedia with Hadoop and Pig for Natural Language Processing(blogs.nuxeo.com)

   51 points by ogrisel 16y ago|2 comments
7. 247.▲
   by ogrisel 16y ago

   AFAIK this is not based on the new architecture introduced in this talk.
8. 248.▲
   by ogrisel 16y ago

   I am full-time R&D engineer at http://nuxeo.com working on integrating semantic analysis technologies into our open-source document management platform and applications. All my code is released under LGPL or ASL2 on either the company
9. 249.▲
   by ogrisel 16y ago

   I agree on this specific case but consistency with conventions shared across projects is more important. But I don't think anybody will complain if you use either of the them whereas 160 chars long expressions with no spacing between operat
10. 250.▲
    by ogrisel 16y ago

    I agree publishing datasets is very very important too. Often more important than code.
11. 251.▲
    by ogrisel 16y ago

    If your system is complicated enough (which can be the case for complex machine learning or NLP algorithms), an 8 pages paper (common limit for many conferences) cannot describe all the implementation details but those implementation detail
12. 252.▲
    by ogrisel 16y ago

    The 80 chars limit has a justification: expressions that don't fit on 80 chars (or two lines using parens) are not readable anyway. In such a case temporary variables with meaningful names would both help respect the 80 chars constraint and
13. 253.▲
    by ogrisel 16y ago

    That must change. Science must be reproducible. Other researchers should be able to dive into each others code quickly to understand the impact of implementation details.
14. 254.▲
    by ogrisel 16y ago

    And the corollary: Why do researchers never respect the PEP8 when they write python code? Yes I am a bit overreacting since the blog post is very well written and I actually agree 100% with the content. But please people: respect the PEP8
15. 255.▲
    by ogrisel 16y ago

    Great talk indeed. Note that most (all?) the material covered in the talk is available from the tutorials referenced on the README of the project on github: http://github.com/nathanmarz/cascalog And the demo project: http://github.com/na
16. 256.▲
    by ogrisel 16y ago

    Yes I know and use theano for some deep learning experiments. This is really a great tool. But theano will probably never be considered a default dependency for common machine learning libs as long as there is no good OpenCL runtime opensou
17. 257.▲
    by ogrisel 16y ago

    Read the NLTK book available online for free here: http://nltk.org The nltk tools them-selves might not always be state of the art (e.g. compare with the java libs from the NLP dept in Stanford) but at least you will get to know all the m
18. 258.▲
    by ogrisel 16y ago

    Learn hadoop / lucene / pig / hive / clojure / python / ruby and put your code on github and if you can digest all of these you will get plenty of interesting job offers.
19. 259.▲
    by ogrisel 16y ago

    xlvector has a blog on recsys that talks about a good but overlooked feature of those systems from a user point of view: the ability to have the user make fortunate discoveries while looking for something seemingly unrelated: + http://xlve
20. 260.▲
    by ogrisel 16y ago

    Thanks for posting this: I read this paper some time ago but completely forgot the name of the authors. It is indeed a good "tricks of the trade" paper. I real enjoy such papers.
21. 261.▲
    by ogrisel 16y ago

    Something that is really missing in my opinion is a good opensource OpenCL [1] runtime for regular x86 CPUs using both muticore optims (a la openmp) and vector instructions (SSE). Bonus points if your OpenCL runtime can also leverage the no
22. 262.▲
    Introducing fise, the Open Source RESTful Semantic Engine(blogs.nuxeo.com)

    3 points by ogrisel 16y ago|0 comments
23. 263.▲
    by ogrisel 16y ago

    Or change your feature extractors :)
24. 264.▲
    by ogrisel 16y ago

    "Related Questions": tokenize => random hash/project (tokens) => TD-IDF => KD-tree lookup "automatically infer tag": tokenize / shingle q&a, ORDER token+bigrams BY TF-IDF(token + bigrams) In both cases a global IDF estimate
25. 265.▲
    by ogrisel 17y ago

    I mostly agree with parent, but the divorce between AI and ML might only be temporary, and the deep learning branch of ML is targeting more general AI than just fitting linear models, c.f. for instance for a high level overview: http://
26. 266.▲
    by ogrisel 17y ago

    money?