Digital Language Resources

DIGITAL LANGUAGE RESOURCES
AND THEIR USE IN EDUCATION
Assist. Prof. Rositsa Dekova, PhD
The Paisii Hilendarsky University of Plovdiv

Department of English Studies
Digital Language Resources
2
 Electronic dictionaries
 Monolingual & Bilingual
 Picture dictionaries
 Learning games
 Annotated Corpora
 BulNC
 BNC
 COCA
 Parallel corpora
 Bulgarian-English Parallel Corpus
 Lexical Semantic Databases
 WordNet
 FrameNet
Educators & Computers
Electronic dictionaries
3
Learner’s dictionaries and thesauri

 http://www.collinsdictionary.com/
 http://www.merriam-webster.com/
 http://thesaurus.com/
 http://www.eurodict.com/
 http://www.lingvozone.com/free-online-dictionary

Picture dictionaries
4
 http://www.oxfordlearnersdictionaries.com/wordli
st/english/pictures/pics_A-B/
 http://www.kidzwood.com/#a-z
 Pictorial dictionary with voice
 http://www.opdome.com/
 Online picture dictionary
 http://www.anglomaniacy.pl/index.html
 Vocabulary & Grammar
 Songs & Printables

Learning English can be fun
5
Crosswords and word searches

 https://learningapps.org/
 https://www.education.com/worksheet-generator/
 http://www.teachers-direct.co.uk/resources/wordsearches/
 http://puzzlemaker.discoveryeducation.com/WordSearchSetupForm.asp?c
ampaign=flyout_teachers_puzzle_wordcross
 http://worksheets.theteacherscorner.net/make-your-own/word-search/
 http://tools.atozteacherstuff.com/word-search-maker/wordsearch.php
 https://www.puzzle-maker.com/wordsearch_Entry
 http://www.armoredpenguin.com/crossword/
 http://www.abcya.com/make_a_word_search.htm

Word search example
6
African Animals
Word search generated by

http://puzzlemaker.discoveryeducation.com

Word search example
7
Word search from the database of

http://www.teachers-direct.co.uk/

Word search example
8

Word search example
9

Interactive games
10
 http://www.eslgamesplus.com/
 http://gamestolearnenglish.com/
 http://www.vocabulary.co.il/english-language-
games/
 http://www.learninggamesforkids.com/vocabulary
_games/foreign-languages.html
 http://www.eslgamesplus.com/
 http://www.abcya.com/
 ...
ANNOTATED CORPORA
11
Corpus – a large body of machine-readable

naturally occurring linguistic evidence.
Annotated Corpus – enhanced with various types
of linguistic information
 Morphological
 POS tagging
 Semantic
 tagging with words senses
 Syntactic
 tagging for syntactic information
 …
TEXT SEGMENTATION
12

TEXT SEGMENTATION
13
 Electronic text is just a sequence of characters.

 Before any processing is done the text has to be
segmented into linguistic units, such as words,
punctuation, numbers, alphanumericals (H2O), etc.
 This process is called TOKENIZATION and the
segmented units are called TOKENS.
 The process of segmenting the text into sentences
is called SENTENCE SPLITTING.

LEMMATIZATION
15
 Reduces inflectional forms and sometimes

derivationally related forms of a word to return
the base or dictionary form of a word, which is
known as the lemma.
 For instance:
 am, are, is  be
 car, cars, car's, cars'  car

PART-OF-SPEECH TAGGING
16
 Parts of speech
 The morphological and syntactic classes that the
different parts of speech can be assigned to.
 POS tagging
 Automatic assignment of descriptors called tags
to input tokens.

THE TAGSET
17
 The tagset includes all the tags that will be

used in the POS tagging.
 We could use a very coarse tagset:
 N, V, Adj, Adv, Prep...
 More commonly used set is finer-grained:

 NN, NNS, NNP, NNPS, VB, VBG, VBN, VBP, VBZ…
 The level of granularity used in the tagset

directly affects the search possibilities.
EXAMPLE
18
 The _ DT
 little _ JJ
 boy_NN1
 quickly_RB
 ate_ VVD
 the_DT
 green_JJ
 apple_NN1
 ./.
CASES OF AMBIGUITY
19
 They love_V summer_Adj vacations.

 Their love_N started in the summer_N.
 Every plant_N needs water and light.

 We should all plant_V at least one tree
in our life.
CASES OF AMBIGUITY
20
 They_PNP  Time_NN1_VVB
 are_VBB  flies_VBZ_NNS
 flying_NN1_VBG  like_PRP_VVB
 planes_NN2  an_AT0
 ./.
 arrow_NN1
Coreference Resolution:
 ./.
Who/what are ‘they’?

THEY ARE FLYING PLANES
21

Examples of Taggers and Parsers
22
 CLAWS WWW tagger (Free web tagging service for English)

 http://ucrel.lancs.ac.uk/claws/trial.html
 The Stanford Parser online

 http://nlp.stanford.edu:8080/parser/
 Shallow Parsing Demo

 Syntactic Tree Generator URL
 An app that builds syntactic trees from labelled
bracket notations.

https://demo.allennlp.org/
23
Constituency Parsing:
Breaks a text into sub-phrases

(constituents).
https://demo.allennlp.org/
24
 Reading comprehension – answers questions

about a passage of text.
 Sentiment Analysis – predicts whether an input is
positive or negative.
 Coreference resolution – finds all expressions that
refer to the same entity in a text.
 Language Modeling – generates the most likely
next words.
 …
Applications for formula reading
25
 www.readspeaker.com
 http://www.robobraille.org/robobraille-projects
 http://www.inftyproject.org/en/index.html
 http://lpf-esi.fe.up.pt/~audiomath/demo/AM1.htm

BRITISH NATIONAL CORPUS (BNC)
26
 A 100 million word collection of samples of

written and spoken language from a wide
range of sources, designed to represent a
wide cross-section of current British English,
both spoken and written.
 Available online at:
https://www.english-corpora.org/bnc/

BRITISH NATIONAL CORPUS (BNC)
27
 The written part of the BNC (90%) includes

 extracts from regional and national newspapers,
 specialist periodicals and journals for all ages and interests,
 academic books and popular fiction, published and unpublished
letters and memoranda,
 school and university essays, etc.
 The spoken part (10%) consists of

 orthographic transcriptions of unscripted informal conversations
(recorded by volunteers selected from different age, region and social
classes in a demographically balanced way)
 spoken language collected in different contexts, ranging from
formal business or government meetings to radio shows and
phone-ins
THE CORPUS OF CONTEMPORARY
AMERICAN ENGLISH (COCA)
28
 The Corpus of Contemporary American

English (COCA) is the largest freely-available
corpus of English, and the only large and
balanced corpus of American English.
 The corpus was created by Mark Davies of
Brigham Young University.
https://www.english-corpora.org/coca/

COCA
29
 Contains 600 million words of text and is

equally divided among spoken, fiction, popular
magazines, newspapers, and academic texts.
 Updated regularly - 20 million words included
each year from 1990-2019.
 Suitable for looking at current, ongoing
changes in the language.

THE COCA SEARCH ENGINE
30
 Searches for exact words or phrases, wildcards,

lemmas, part of speech, or any combinations of
these.
 Searches for surrounding words (collocates)
within a ten-word window.
 Limit searches by frequency and compare the
frequency of words, phrases, and grammatical
constructions:
 bygenre or even between sub-genres (or domains)
 over time

Results for collocates of black
31

SEMANTICALLY-BASED QUERIES OF THE CORPUS
32
 Contrast and compare the collocates of two related

words (little/small, democrats/republicans,
men/women).
 Determine the difference in meaning or use between
these words.
 Find the frequency and distribution of synonyms for
nearly 60,000 words
 Compare the frequency in different genres.
 Create your own lists of semantically-related words, and
then use them directly as part of the query
SPECIFIC USES OF COCA
33
 To look at recent changes in English:

 morphology (new suffixes –friendly and –gate)
 syntax (including prescriptive rules, quotative like, so
not ADJ, the get passive, resultatives, and verb
complementation)
 semantics (such as changes in meaning with web,
green, or gay)
 lexis – including word and phrase frequency by year,
to produce lists of all words that have had large shifts
in frequency all words that have had large shifts in
frequency between specific historical periods.
PARALLEL CORPORA
34
 Alignment (of bitexts)

 Differences in grammatical structure
 with the sun not shining - нямаше слънце
 Differences in lexical structure
 the thermometer walks inch by inch up to the top of the
glass,  и термометърът пълзи сантиметър по
сантиметър до върха на скалатa
 No lexicalization
 It wasn't the butler coming back.  Не беше икономът.
 It’s this way  Положението е такова
INTELLIGENT SEARCHES IN BG-EN PARALLEL CORPUS
35
 The Bulgarian National Corpus search engine is available at:

http://search.dcl.bas.bg/
 The syntax allows search by (combinations of) word forms,
grammatical tags, semantic relations.
 Thanks to the alignment, the corresponding sentences in
parallel documents are also accessible.
 The hits are paginated and the matches are highlighted.
 The user is able to view the detailed information for a given
sentence in the hit set - the sentence metadata, its context,
and correspondence(s) in the other languages.
SEARCH ASSISTANT
36

LEXICAL SEMANTIC NETWORKS
37
Electronic language resources which define notions

through their relations with other notions.
 LSN are knowledge representation schemes involving nodes
and links (arcs or arrows) between nodes.
 The nodes represent objects or concepts.
 The links represent relations between nodes.
 The links are directed and labeled.

An example of classical taxonomy tree
38
boy = male child

girl = female child
man = male adult
woman = female adult
child = young human human
adult = grown-up human
adult child
[+adult] [-adult]
man woman boy girl

[+male] [-male] [+male] [-male]

Lattice structure with multiple classifications
39

WORDNET - http://wordnet.princeton.edu/
40
 A large lexical semantic database of English

 Originally developed at Princeton University (Miller, 1990)
 EuroWordNet - http://www.illc.uva.nl/EuroWordNet/
 BalkaNet - http://www.dblab.upatras.gr/balkanet/index.htm
 Each wordnet represents a unique language-internal
system of lexicalizations
 In addition, the wordnets are linked to an Inter-Lingual-
Index, based on the Princeton wordnet

41
WORDNET STRUCTURE
 Nouns, verbs, adjectives and adverbs are grouped
into sets of cognitive synonyms (synsets), each
expressing a distinct concept.
 Each synset is linked to other synsets by means of a
small number of “conceptual relations.”
 WordNet really consists of four sub-nets, one each
for nouns, verbs, adjectives and adverbs, with few
cross-POS pointers.

42
WORDNET STRUCTURE
http://wordnet.princeton.edu/man/wngloss.7WN.html
 Each synonym set - SYNSET - encodes the relation of
equivalence between a number of lexical items –
LITERALS where each lexeme:
 has unique meaning (specified by the value of SENSE)
 pertains to one and the same part of speech
(specified as the value of POS)
 represents one and the same lexical meaning
(specified as the value of DEF - definition)
43

An example: learn (Wordnet)
44

BulNet http://dcl.bas.bg/bulnet/
45
 A lexical semantic network of Bulgarian

 comprises around 49,189 synonym sets
distributed into nine parts of speech
 open-class words: nouns, verbs, adjectives and
adverbs
 closed-class words: pronouns, prepositions,
conjunctions, particles and interjections

46
STRUCTURE
 Each synset is linked to its counterpart in PWN3.0 by
means of a unique identification number – ID.
 The common synsets in the Balkan languages are
marked as common concepts subsets – BCS.
 In the monlingual database a synset should be linked to
at least one other synset through an intralingual
relation.
 Non-obligatory information may also be encoded such
as examples of usage, stylistic, morphological or
syntactic properties.
RELATIONS IN BULNET
47
Synonymous sets are linked through various relations:

 SEMANTIC
 Synonymy, antonymy, hypernymy, hyponymy, meronymy,
holonymy, entailment, inclusion, causation, etc.
 MORPHOSEMANTIC
 BE IN STATE
 MORPHOLOGICAL
 DERIVED
 PARTICLE
 EXTRALINGUISTIC
SEMANTIC RELATIONS
48
SYNONYMY – a semantic relation of equivalence

between literals belonging to the same POS;
The synonyms form the synonym set also called
SYNSET.
For example:
 The lexical units
{auto:1, car:2, automobile:2, machine:3, motorcar:1}
form a synset as they refer to the same concept.
SEMANTIC RELATIONS
49
HYPERNYMY and HYPONYMY - semantic relations between

synsets, which corresponds to the notion of class-
inclusion: if W1 is a kind of W2, then W2 is
hypernym of W1 and W1 is hyponym of W2.
Example:
 rose < plant < living organism
Multi-parent relations:
 actress < actor
 actress < female.
SEMANTIC RELATIONS
50
ANTONYMY – a semantic relation of opposition,

established between two members belonging to
one and the same POS.
Examples:
 man - woman
Hyponyms of two antonyms (nouns) should also be
antonymous pair by pair:
 man - woman
 actor - actress
SEMANTIC RELATIONS
51
MERONYMY and HOLONYMY – semantic relations linking

synsets denoting wholes with those denoting their
parts: if W1 has a W2, and W2 is part, portion,
member of W1, then W1 is a meronym of W2 and
W2 is a holonym of W1
Examples:
 ...
MERONYMY may not be always reversible to HOLONYMY:
 tree - forest
TYPES OF MERONYMY
52
 PART OF:
 клон – дърво
 книга - библиотека
 MEMBER OF:
 дърво – гора
 футоболист – отбор – лига
 PORTION OF:
 капка – течност
WORD-FORMING RELATIONS
53
Morpho-semantic relations
 BE IN STATE
 кост – костен
Morphological relations
 DERIVED
 сервирам – сервитьор
 PARTICIPLE
 видя – видян

EXTRALINGUISTIC RELATIONS
54
 REGION DOMAIN
 степ – Русия
 USAGE DOMAIN
 тиранти – множествено число
 CATEGORY DOMAIN
 гимнастически уред – гимнастика
THE RELATIONS IN BULNET
55
The large number of relations encoded in the

Bulgarian wordnet effectively illustrates the
language's semantic and derivational
richness
This offers diverse opportunities for numerous

applications of the multilingual database.

56

57

58

APPLICATIONS
59
 options for synonym selection

 queries for semantic relations of a word in the
language's lexical system
 antonymy, holonymy, etc.
 explanatory definition queries

 translation equivalents for a lexical item

FRAMENET (Fillmore and Baker 2001, 2010)
60
 A lexical database of English that is both human-

and machine-readable.
 Based on annotated examples of how words are
used in actual texts.
 Tries to capture human insight into how a word
can be used and converts it into semantic
knowledge that is machine-readable.
http://www.icsi.berkeley.edu/~framenet
FRAME SEMANTICS (Fillmore, 1976, 1985)
61
 A semantic frame is a structure used to define the

semantic meaning of a word.
 Cutting
 Frame elements are the separate elements which

make up a frame.
 An Agent cuts an Item into Pieces using an Instrument.
 Lexical units are the words that evoke a particular
frame.
 carve.v, chop.v, cube.v, cut.v, dice.v, fillet.v, mince.v,
pare.v, slice.v

62 Educators & Computers
OTHER TOOLS
65
 Quizlet - https://quizlet.com/
 Padlet - https://bg.padlet.com/
 Trello - https://trello.com/create-first-board

Uses of Digital Language Resources
66
 EDUCATION
 Intelligentsearches for particular language
phenomena, i.e. search by (combinations of) word
forms, grammatical tags, semantic relations;
 Collocations;
 Word and phrase frequencies;
 Recent changes in the language;
 Translation equivalents;
 Semantic structure of the words and their use;
 etc.
References
67
 Davies, Mark. 2010. The Corpus of Contemporary American

English as the first reliable monitor corpus of English Lit
Linguist Computing (2010) 25 (4): 447-464 first published
online October 27, 2010 .
 The British National Corpus, version 3 (BNC XML Edition).
2007. Distributed by Oxford University Computing Services on
behalf of the BNC Consortium. URL:
http://www.natcorp.ox.ac.uk/
 Reference Guide for the British National Corpus (XML Edition)
edited by Lou Burnard, February 2007. URL:
http://www.natcorp.ox.ac.uk/XMLedition/URG/
References
68
 Miller, George A. (1995). WordNet: A Lexical Database for English.

Communications of the ACM Vol. 38, No. 11: 39-41.
 Fellbaum, Christiane (1998, ed.) WordNet: An Electronic Lexical
Database. Cambridge, MA: MIT Press.
 Koeva, S., T. Tinchev and S. Mihov. Bulgarian Wordnet - structure and
validation. In Romanian Journal of Information Science and
Technology, Vol. 7, No. 1-2, 61-78, 2004. ISSN 1453-8245 pdf file
 Koeva, S. Derivational and morphosemantic relations in Bulgarian
Wordnet. In Intelligent Information Systems, XVI, Warsaw, Academic
Publishing House, 2008, 359-389. ISBN 978-93-60434-44-4 pdf file

References
69
 Ruppenhofer, J. et al. 2010. FrameNet II: Extended Theory and

Practice. https://framenet2.icsi.berkeley.edu/docs/r1.5/book.pdf
 Fillmore, Charles. Introduction to FrameNet.
https://framenet.icsi.berkeley.edu/fndrupal/sites/default/file
s/FNintroCJF.ppt
 Fillmore, Charles J. 1985. Frames and the Semantics of
Understanding. Quaderni di Semantica 6(2): pp. 222-53.

THANK YOU
FOR YOUR ATTENTION!

Digital Language Resources

Uploaded by

Copyright:

Available Formats

Digital Language Resources

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Digital Language Resources

Uploaded by

Copyright:

Available Formats

DIGITAL LANGUAGE RESOURCES

AND THEIR USE IN EDUCATION

Assist. Prof. Rositsa Dekova, PhD

The Paisii Hilendarsky University of Plovdiv

Learner’s dictionaries and thesauri

Educators & Computers

Educators & Computers

Crosswords and word searches

Educators & Computers

Word search generated by

Educators & Computers

Word search from the database of

Educators & Computers

Educators & Computers

Educators & Computers

Corpus – a large body of machine-readable

Educators & Computers

 Electronic text is just a sequence of characters.

Educators & Computers

 Reduces inflectional forms and sometimes

Educators & Computers

Educators & Computers

 The tagset includes all the tags that will be

 More commonly used set is finer-grained:

 The level of granularity used in the tagset

 They love_V summer_Adj vacations.

 Every plant_N needs water and light.

Educators & Computers

Educators & Computers

 CLAWS WWW tagger (Free web tagging service for English)

 The Stanford Parser online

 Shallow Parsing Demo

Educators & Computers

Breaks a text into sub-phrases

 Reading comprehension – answers questions

Educators & Computers

 A 100 million word collection of samples of

Educators & Computers

 The written part of the BNC (90%) includes

 The spoken part (10%) consists of

 The Corpus of Contemporary American

Educators & Computers

 Contains 600 million words of text and is

Educators & Computers

 Searches for exact words or phrases, wildcards,

Educators & Computers

Educators & Computers

 Contrast and compare the collocates of two related

 To look at recent changes in English:

 Alignment (of bitexts)

 The Bulgarian National Corpus search engine is available at:

Educators & Computers

Electronic language resources which define notions

Educators & Computers

boy = male child

man woman boy girl

Educators & Computers

Educators & Computers

 A large lexical semantic database of English