Archives: Tweets
-
#nwnlp16 Johnson et al. are extending ConVis https…
#nwnlp16 Johnson et al. are extending ConVis https://t.co/hEqJnGIVt1 and suggest it might be useful for journalists visiting Reddit &c
-
#nwnlp16 and @pyry’s not even here! https://t.co/J…
#nwnlp16 and @pyry’s not even here! https://t.co/JwQZAW0qED
-
#nwnlp16 discourse pairs have “very high probabili…
#nwnlp16 discourse pairs have “very high probabilities of discourse connectivity” in ‘elaboration’; ‘elaboration’ is Bayesian wastebasket
-
#nwnlp16 doing rhetorical parsing as “a pseudomono…
#nwnlp16 doing rhetorical parsing as “a pseudomonologue” — many conversations are true monologues, especially when s.o. like me has 1 half
-
#nwnlp16 you know how you have 2 conversations in…
#nwnlp16 you know how you have 2 conversations in one thread? … yeah, that’s what we’re doing here.
-
#nwnlp16 Linking Sentences in Asynchronous Convers…
#nwnlp16 Linking Sentences in Asynchronous Conversations, Jordon Johnson, Giuseppe Carenini and Raymond Ng presented by Jordon
-
RT @rctatman: Ammar et al: https://t.co/7ZLrRH6HV8…
Ammar et al: https://t.co/7ZLrRH6HV8 online evaluation portal using same tests we did #nwnlp16
-

#nwnlp16 two very different posters both concerned…
#nwnlp16 two very different posters both concerned with latency/quality tradeoffs for high level NLP (MT & parsing). https://t.co/kRSKi8iMqB
-
#nwnlp16 Ammar &c evaluation QVEC-CCA supports…
#nwnlp16 Ammar &c evaluation QVEC-CCA supports evaluation WRT tag set; argue that tagset is available across langs & thus better metric
-
#nwnlp16 also do multilingual clustering as useful…
#nwnlp16 also do multilingual clustering as useful replacement to monolingual corpus for translation. Evaluation hard, proposes new method