Twitter | Search | |
Search Refresh
Graham Neubig 2h
internet, what are seminal examples of "style transfer" tasks: paraphrasing with a stylistic target? I know text simplification (), spoken->written language (), "paraphrasing for style" (). Any others?
Reply Retweet Like
Ryan D. Cotterell Sep 23
A pattern such as the Hearst pattern allows one to extract hypernymy easily with simple regular expressions in an unsupervised fashion. The original idea (Hearst 1992) [] is still used today when text-mining from large corpora.
Reply Retweet Like
Daniil Sorokin Sep 17
Replying to @daniilmagpie
The background: we are looking for data sets that involve factual (world) knowledge.
Reply Retweet Like
Tim Baldwin 2h
videos are now up for all oral presentations: . Let me know if there are any issues.
Reply Retweet Like
Gerard de Melo Sep 18
Controversial claims about many "so-called" experts. Agree or disagree with views expressed in this CACM blog post by W. Saba?
Reply Retweet Like
Jingbo Shang Sep 18
No line-by-line annotations -- train named entity taggers with distant supervision and get competitive performance. 🚀🚀🚀
Reply Retweet Like
Julia Rozanova Sep 14
Big highlight for enthusiasts yesterday at the : succinct field overview by and many insights from the panel discussion on open problems in NLP. With , , , and
Reply Retweet Like
Kirk Borne Sep 22
Reply Retweet Like
Kirk Borne Sep 18
How machines understand our language — an introduction to Natural Language Processing:
Reply Retweet Like
José Pérez Moreno Sep 19
Trying to train (pytorch) with English-Croatian parallel texts. I am certainly sure both files have the same number of lines, but I keep getting the same error. I guess it must be some weird Unicode character somewhere 😥
Reply Retweet Like
Gabriela Czarnek 18h
So let's find out! How can we do this using existing data, e.g., Twitter? Any ideas from wizards?
Reply Retweet Like
Isabelle Augenstein 9h
One week left to apply for a fully funded PhD position!
Reply Retweet Like
Kirk Borne 21h
Reply Retweet Like
vachagan gratian Sep 23
Hello World Here is a program to detect emotions of tweets (or short texts for that matter)
Reply Retweet Like
Jose Camacho Collados Sep 17
Great to see more and more datasets of standard tasks available in many different languages!
Reply Retweet Like
Tariq ☕ Sep 20
which of these word clouds is easier to digest and work out what the underlying text is about (fairytales)? you might be surprised at the algorithm
Reply Retweet Like
Siobhán Grayson Sep 22
Word embeddings from "A Minimal Turing Test" by JP McCoy & Paper: Select one word from the English dictionary to prove that you're human. "Love" was the most popular choice. notebooks :
Reply Retweet Like
cs.CL Papers Sep 20
A Quantitative Evaluation of Natural Language Question Interpretation for Question Answering Systems. (arXiv:1809.07485v1 [])
Reply Retweet Like
Tariq ☕ 45m
Reply Retweet Like
Kirk Borne 21h
Reply Retweet Like