I am moving code and project folders to GitHub. I don’t know, whether this is a good idea, it just turns out to be easier to use…

This port includes the SNLTK code, all kinds of Python 3 projects, Java code, some of the C(++) code for FSTs and some NLP tasks, corpus and TEI XML utils. Some of that I limited to pull-only and push-access exclusively for collaborators. If you were involved in some of that, let me know, send me your GitHub-ID is and I can add you to the collaborators group of the particular repos.

In particular, my course material will be migrated to GitHub completely. For example, the course material for the LSA Summer Institute course in summer 2013 will be placed there:
Python 3 for Linguists

Read More...

Tags: GitHub, Python3, Scheme, Java, Course material, C

Since I was asked many times about this MA program and the original text that went to the accreditation committee in Croatia (where we got one very nasty and absolutely irrelevant review, if I find it, I’ll post it here; but also a very good and constructive review), here are the files, the Croatian and English text about the MA program in Linguistics that we submitted for accreditation within the Bologna system back in 2008 at the University of Zadar. I think, this is the corrected version. It was not the best possible program, developed under time pressure and in a very difficult situation, and its was building on the growing wave of computational linguistics, speech and language technology, as well as theoretical linguistics. We would do a lot of things differently nowadays. If you can use any of this for your inspiration or personal attempts to apply for a program or other support, let us know. I can forward you the editable version for some Office package.

English version:

• Part 1
• Part 2
• Part 3

• Croatian version:

• Part 1
• Part 2
• Part 3

Read More...

Tags: University of Zadar, Linguistics, MA, Program, Bologna

If you want to batch convert a lot of files to some more accessible format (for example ODT or DOCX to HTML or TEI XML), you can use first of all LibreOffice.

Here is a brief introduction how to batch convert files to some LibreOffice output format or TEI XML.
Read More...

Tags: LibreOffice, XSLT, TEI, Corpus

If you write out a stack (or network) in XFST to a prolog file:

write prolog > mymorph.plg

and you want to convert it to DOT and visualize it in Graphviz, here is a Python 3.x script to do so:

Download zipped Python source
View Python code

Read More...

Tags: XFST, DOT, Python

I’ll give a talk on corpora and relevant technologies at Wayne State University in Detroit on the 19^th of October at 11 AM. Here are some links, papers and slides that might be interesting for colleagues and students to follow and post process:

Read More...

Tags: Corpora, WSU, Philologic, LLC, Gutenberg Project, XML, TEI, Lucene, CoreNLP, JavaScript

Date: April 13th, 2012
Time: 1:30 PM
Location: Cooper Building, Suite 104, EMU, 2000 Huron River Drive, Ypsilanti

Directions: Take Washtenaw heading east from Ann Arbor toward Ypsilanti. Go past Hwy 23, turn left on Golfside, then turn right on Huron River Drive. The Cooper Building will be on the left, across from Rynearson Stadium, and there is free parking right out front. If you reach Superior St. you have gone too far.

Title: On the influence of L1 on the L2 perception: The case of tenseness contrast in American vowels
Author: Malgorzata E. Cavar

Abstract:
One obvious difficulty in foreign language learning is the production of foreign sounds. What is less obvious is the fact that the perception of foreign categories by L2 learners differs from that of the native speakers and in itself might be and often is a hurdle in the acquisition of the phonetic/phonological system of the foreign language. In this talk, I will present the results of a series of experiments pertaining to the perception of the English vocalic contrast in high vowels by learners with different L1 backgrounds. The goal of this and similar studies is to determine how perceptual strategies of L2 learners differ from those of English native speakers and what these differences depend on. In the long run, the aim is to predict “customized” areas of difficulty for learners with different backgrounds and to help develop curricula and teaching aids that would actually respond to learners’ needs.
Read More...

Tags: Talk, LINGUIST List, ILIT, Cooper Building, Phonetics, Phonology, EMU

This is a brief description about how to use the Python 3 scripts to generate N-gram models for word tokens and characters from text. I expect you to have a Python 3 interpreter installed on your system.

Read More...

Tags: Python3, N-Gram Model, Frequency profile, Py3k, Tokenization, Tokenizer

The LINGUIST List corpora can be found here:

http://ltl.emich.edu/llc/

You can find in there the LINGUIST List mailings converted to TEI P5 XML. The linguistically annotated version will be available in an extended interface.

See the previous blog for instructions on how to use Philologic…

Read More...

Tags: LINGUIST List, Corpus, Corpus linguistics, Books, Diss

Here is a brief first introduction to the Philologic interface for the LTL corpora and the LINGUIST List corpus;

Read More...

Tags: LTL, Corpora, Corpus, Corpus linguistics, Philologic

I have been asked by many students and colleagues, how to generate nice looking trees for presentations, assignments, papers etc. Here is a small summary of tools I have tried or seen.

If you want to generate a graph of a syntactic relation, a syntactic tree, there are various ways to do that, without manually drawing it on paper and scanning the manual work... here is a small summary of ways and tools for generating syntactic trees...
Read More...

Tags: Trees, Syntax, LFG, AVM, LaTeX, LaTeXIt, Qtree, TikZ

The first version of the small LTL corpus with a couple of million tokens is online. It contains TEI P5 XML encoded books from the public domain. See here…
Read More...

Tags: LTL, Language Technology Lab, EMU, Corpus, English

The online OxGarage Converter on the TEI pages converts almost anything to something else, in particular to TEI XML. This is obviously using the OpenOffice filters and converters in the backend as batch processors, as described here for the manual conversion.
Read More...

Tags: TEI, XML, OpenOffice, Conversion

The poster for the DGfS annual meeting 2012 on a Lithuanian Morphology and LFG Grammar is done. This was the result of a grad course at the University of Konstanz on rule-based natural language processing (using XFST and XLE). I am proud of all the participants!
Here is the poster. You can test the morphology online. The coverage will improve, this is based on the morpheme numbers from the poster, without generic morphological rules. The generator will be made available there too.
Read More...

Tags: LFG, XLE, XFST, Lithuanian, Morphology, Parser, Syntax, DGfS

The LINGUIST List store on amazon.com…
Read More...

Tags: LINGUIST List, Amazon, Books

Please consider supporting LINGUIST List, just go to the Fund Drive 2012 pages and donate!
Read More...

Tags: LINGUIST List, Fund Drive

I set up a simple testing page for a wrapper of raw text to TEI XML. It uses in this version just the Stanford CoreNLP tools to tokenize, recognize sentences, part of speech annotate and lemmatize the input. Just paste a paragraph of text in there. In the next version this will be expanded with NLP tools for a couple of more languages, as well as other analysis components and tools for English.

Read More...

Tags: TEI, XML, CoreNLP, Parsing, Tagging, GATE, UIMA, Syntax, Named entity recognition

Ben Cool ported Charty (CFG-based Chart parser) to JavaScript for a class project and added in one version feature augmentation and unification to it. You can test it online. This is running on mobile devices like iPad or iPhone in Safari and on Android with a browser that has JavaScript support without any server-based component. See the documentation and test site here…

Read More...

Tags: Charty, Syntax, Parsing, JavaScript

To make the Stanford CoreNLP tools work on your Mac OS X 10.7.x (Lion) distribution with the included bash script do this...

Read More...

Tags: CoreNLP, Stanford, corenlp.sh

You should really consider joining this LREC 2012 workshop on Challenges in the management of large corpora!

Read More...

Tags: LREC, Large corpora

Since privacy policy changes seem to be all around now, here is one by me for the pages here:

If you want to make your web-experience somewhat more private, and prevent me from being able to read out something from the apache log files about you, here are some hints about how you could configure your browser to reduce the amount of personal bits you leave on your way on this page or anywhere else on the web:

Read More...

Tags: Privacy, Firefox, DuckDuckGo, Tracking, AdBlock Plus, Ghostery, NoScript, EFF, Panopticlick

The Language Technology Lab (LTL) (ILIT and EMU) is up, check it out:

http://ltl.emich.edu/

More content to come in the next days and weeks… stay tuned!

Read More...

Tags: LTL, Web pages, EMU

Here is a short instruction on using Antconc for simple statistical analysis.

Read More...

Tags: Antconc, Frequency profile, Bigrams, Unigrams, Corpus, KWIC

Here are some of the dictionaries for the OS X Dictionary.app:

• The dict.cc dictionary plugin English-German, German-English
• Tekl.de German Thesaurus and English-German dictionary

Read More...

Tags: Dictionaries, OS X, German, English

The TikZ-dependency graph library for LaTeX can be found here…
Read More...

Tags: Dependency graphs, LaTeX

Here is an online tool for IPA transcription, i2speak:

http://www.i2speak.com/
Read More...

Tags: IPA

I just restored a bunch of web pages of summer schools and workshops. Some had interesting material on them, in particular pictures. Check out the JSSECL 2006 event…

• Fourth Annual Meeting of the Slavic Linguistic Society SLS 2009
• Student Conference on Empirical and Computational Linguistics (JSSECL WS CECL 2006), Zadar, Croatia
• Workshop on Computational Modeling of Lexical Acquisition (CPALA 2005), Split, Croatia
• BOOT-LA workshop at Indiana University

• Computational Linguistics Summer School 2010 at the University of Zadar (CLS2010) (see on Facebook)
• Jadertina Summer School in Empirical and Computational Linguistics (JSSECL 2006)

Read More...

Tags: Summer School, JSSECL, CLS2010, CPALA2005, SLS2009, Web pages

Check out these demonstrations from the Wolfram Demonstrations Project:

• Collocation by Chi Square
• Collocation by Symmetric Conditional Probability

• Multilanguage Word Lengths

• Zipf's Law Applied to Word and Letter Frequencies

• and all the other Linguistic Demonstrations there...

Read More...

Tags: Wolfram, Linguistics, Chi Square, Chi2, Collocation, Zipf's Law, Symmetric Conditional Probability

You should submit a paper to Computational Formal Approaches to Slavic Languages (C-FASL) 2012:

http://cl.indiana.edu/~cfasl/

Read More...

Tags: Slavic, Corpus linguistics, Computational Linguistics, Conference