archive-edu.com » EDU » C » COLUMBIA.EDU

Total: 442

Choose link from "Titles, links and description words view":

Or switch to "Titles and links view".

  • introducing a factor of max i a i 2 b i 2 b i 2 a i 2 in the bound which ultimately leads to a significantly worse result Note There is no problem if A B Thanks to Joel Tropp for pointing out this bug A slightly different algorithm avoids this issue and leads to the same running time bound see http arxiv org abs 1211 5414 for details

    Original URL path: http://www.cs.columbia.edu/~djhsu/papers/randmatrix-errata.txt (2016-02-17)
    Open archived version from archive



  • with very large or infinite strategy sets 1 2 1 R Kleinberg Online decision problems with large strategy sets Ph D Thesis MIT 2005 2 R Kleinberg Anytime algorithms for multi armed bandit problems In Proceedings of the 17th ACM

    Original URL path: http://www.cs.columbia.edu/~djhsu/papers/nhedge-note.txt (2016-02-17)
    Open archived version from archive

  • Tensor Decompositions for Learning Latent Variable Models
    of moments focusing on algorithms based on low rank decompositions of higher order tensors The target audiences of the tutorial include i users of latent variable models in applications and ii researchers developing techniques for learning latent variable models The only prior knowledge expected of the audience is a familiarity with simple latent variable models e g mixtures of Gaussians and rudimentary linear algebra and probability The audience will learn

    Original URL path: http://www.cs.columbia.edu/~djhsu/tutorials/aaai2014/ (2016-02-17)
    Open archived version from archive

  • Tensor Decomposition Methods for Latent Variable Model Estimation
    the method of moments focusing on algorithms based on low rank decompositions of higher order tensors The target audiences of the tutorial include i users of latent variable models in applications and ii researchers developing techniques for learning latent variable models The only prior knowledge expected of the audience is a familiarity with simple latent variable models e g mixtures of Gaussians and rudimentary linear algebra and probability The audience

    Original URL path: http://www.cs.columbia.edu/~djhsu/tutorials/icml2013/ (2016-02-17)
    Open archived version from archive

  • Daniel J. Hsu | Information for prospective students
    me you can mention this in your application and also see the section below this is likely the best way to draw my and others attention to your file Note that there are several faculty members at Columbia working on machine learning theory of computation and foundations of data science Unfortunately I do not generally have time to reply to inquiries from prospective students Already at Columbia and interested in

    Original URL path: http://www.cs.columbia.edu/~djhsu/prospective.html (2016-02-17)
    Open archived version from archive

  • Non-gross
    Backstory Cafe Medici on 57th Philadelphia PA Elixr Coffee La Colombe The Ranstead Room Spruce Street Espresso Food and Friends Garces Trading Company KoJa Nan Zhou Hand Drawn Noodle House Pub Kitchen New York NY Abraço Burp Castle Cafe Grumpy Culture Espresso Bar Double Dutch Espresso Irving Farm Coffee Company Joe Coffee Ninth Street Espresso Nous Espresso Bar Prodigy Coffee Caracas Arepa Bar Celeste Jin Ramen Katsu Hama Klong Larb

    Original URL path: http://www.cs.columbia.edu/~djhsu/endorse.html (2016-02-17)
    Open archived version from archive


  • column is the original text Requires enscript and ps2pdf I suggest trying this with texts from http www gutenberg org For more about bkoo see http bkoo org use strict sub print pages my lines ref fh num lines foreach lines ref print fh join map join sort lc a cmp lc b split substr 0 1 substr 1 1 split push orig lines if lines per page 0 print

    Original URL path: http://www.cs.columbia.edu/~djhsu/code/bkoo (2016-02-17)
    Open archived version from archive


  • fi if use wide eq 1 then slide w 1920 slide h 1080 else slide w 1024 slide h 768 fi set temp directory tmpdir tmp p2k basename 2 if d tmpdir then echo Error temp directory tmpdir already exists please delete it first dev stderr exit 1 fi mkdir tmpdir extract pages using pdftk pdftk 1 burst output tmpdir d pdf determine number of pages if e tmpdir doc data txt then numpages grep NumberOfPages tmpdir doc data txt cut d f 2 rm f tmpdir doc data txt else echo Error pdftk did not generate doc data txt dev stderr exit 1 fi jump to temp directory pushd tmpdir dev null create index file cat index apxl End of header for page in seq numpages do if e page pdf then echo Error pdftk did not generate page pdf dev stderr exit 1 fi determine scaling and position if has pdfinfo eq 1 then pdf w pdfinfo page pdf grep Page size tr s cut d f 3 pdf h pdfinfo page pdf grep Page size tr s cut d f 5 fi if echo scale 10 slide w slide h pdf w pdf h bc eq 1

    Original URL path: http://www.cs.columbia.edu/~djhsu/code/pdf2keynote (2016-02-17)
    Open archived version from archive