Changes between Version 5 and Version 6 of private/AdvancedNlpCourse/LanguageModelling


Ignore:
Timestamp:
Nov 2, 2015, 7:40:51 AM (5 years ago)
Author:
Vít Baisa
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • private/AdvancedNlpCourse/LanguageModelling

    v5 v6  
    1818== Practical Session ==
    1919
    20 We will build a simple language model (skip-gram) which has very interesting properties. When trained properly, the vectors of words obey simple space arithmetics, e.g.
    21 vector "king" − vector "man" + vector "woman" ~= vector "queen".
    22 We will train this model on a large Czech and English corpora and evaluate the results.
     20We will build a simple character-based language model and generate naturally-looking sentences. We need a plain text and fast suffix sorting algorithm (mksary).
     21
     22
     23
     24
     25=== Task ===
     26
     27Change the training process and the generating process to generate the most naturally-looking sentences. Either by
     28* pre-processing the input plain text or
     29* setting training parameters or
     30* changing generating process
     31* or all above.