+ All Categories
Home > Technology > Digital textualities news2lyric

Digital textualities news2lyric

Date post: 21-Jan-2017
Category:
Upload: anneleen-artois
View: 77 times
Download: 0 times
Share this document with a friend
12
Digital Textualities News2Lyric B-KUL- G0R06A Anneleen Artois
Transcript
Page 1: Digital textualities   news2lyric

Digital TextualitiesNews2Lyric

B-KUL-G0R06AAnneleen Artois

Page 2: Digital textualities   news2lyric

Structure

Inspiration

News2LyricFunctionality

Code

Demonstration

Future Research

Page 3: Digital textualities   news2lyric

Part I : Inspiration

Generative poetry is produced by programming algorithms and drawing from corpora to create poetic lines. This is the oldest e-poetic genre and remains relevant today through e-literary genres.

Interactive poetry incorporates input from the reader in the e-poem’s expressive strategies.

(iloveepoetry.com, Flores)

(iloveepoetry.com, Flores)

Born-Digital Literature: Digital Poetry

Page 4: Digital textualities   news2lyric

Nick Monfort - Taroko Gorge (2009)

Page 5: Digital textualities   news2lyric

Part II : News2Lyric

- Rhyme - Meter- Stress - Occurrence of certain words :

- Subject- Verbs- Content words ( = no stop

words )> Input = title

Page 6: Digital textualities   news2lyric

Structure- ACL 2014 Ninth Workshop on Statistical Machine

Translation: News Crawl: articles from 2007

guinea_pig = ( [[False, True, False, True, False, True, False, True], [False, True, False, True, False, True, False, True], [False, True, False, True, False, True, False, True], [False, True, False, True, False, True, False, True]], [1, 1, 2, 2],)

Page 7: Digital textualities   news2lyric

Jupyter Notebook

Page 8: Digital textualities   news2lyric

A

Recapitulate…

CorpusB

Title

Article : Body

API / Text File

?

PoS + Dependency Parsing

from nltk.parse.stanford import StanfordDependencyParserfrom nltk.stem.porter import PorterStemmer“This course has inspired me generating poetry”

[((u'thought', u'VBN'), u'nsubj', (u'course', u'NN')), ((u'course', u'NN'), u'det', (u'This', u'DT')), ((u'thought', u'VBN'), u'aux', (u'has', u'VBZ')), ((u'thought', u'VBN'), u'dep', (u'reading', u'VBG')), ((u'reading', u'VBG'), u'nsubj', (u'me', u'PRP')), ((u'reading', u'VBG'), u'dobj', (u'poetry', u'NN'))]

> subject: course, verb: inspired, verb: generating> stop words

Title = Title Poem

from nltk.corpus import cmudictget_phonemes("hello"): [[u'HH', u'AH0', u'L', u'OW1'], [u'HH', u'EH0', u'L', u'OW1']]et_phonemes("world"): [[u'W', u'ER1', u'L', u'D']]

http://localhost:5000/

Web framework

Page 9: Digital textualities   news2lyric

Demonstration

Page 10: Digital textualities   news2lyric
Page 11: Digital textualities   news2lyric

Part III : Future Research

● Add Different Poem Structures● API : Titles + Links to

Articles● Neural Net : Line + Phonetic

Representation● Topic Modeling : LDA (?)● Current Topics● More Interactive Interface

Page 12: Digital textualities   news2lyric

Questions ?


Recommended