BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Internal Seminar  - NLIP Postdocs and PhDs
DTSTART:20151023T110000Z
DTEND:20151023T120000Z
UID:TALK61213@talks.cam.ac.uk
CONTACT:Tamara Polajnar
DESCRIPTION:This seminar is open to members of the NLIP group and MPhil st
 udents doing NLP projects in the Computer Lab. It features 4 talks:\n\nAn 
 Exploration of Discourse-Based Sentence Spaces for Compositional Distribut
 ional Semantics \n\nTamara Polajnar and Laura Rimell\n\nWe investigate whe
 ther the wider context in which a sentence is located can contribute to a 
 distributional representation of sentence meaning. We compare a vector spa
 ce for sentences in which the features are words occurring within the sent
 ence\, with two new vector spaces that only make use of surrounding contex
 t. Experiments on simple subject-verb-object similarity tasks show that al
 l sentence spaces produce results that are comparable with previous work. 
 However\, qualitative analysis and user experiments indicate that extra-se
 ntential contexts capture more diverse\, yet topically coherent informatio
 n.\n\n\n-----------------------------\n\nWhat Happens Next? Suggesting Eve
 nts in Narratives using Neural Event Representations\n\nMark Granroth-Wild
 ing\n\n\n-----------------------------\n\nOnline Representation Learning i
 n Recurrent Neural Language Models\n\nMarek Rei\n\nThis work extends recur
 rent neural network language models with continuous online learning. The m
 odel keeps a separate vector representation of the current unit of text be
 ing processed and adaptively adjusts it after each prediction. The initial
  experiments give promising results\, indicating that the method is able t
 o increase language modelling accuracy\, while also decreasing the paramet
 ers needed to store the model along with the computation required at each 
 step.\n\n\n-----------------------------\n\nWhat happens when you encode a
  dictionary or encyclopedia in a neural net\n\nFelix Hill\n\n\n
LOCATION:FW26\, Computer Laboratory
END:VEVENT
END:VCALENDAR
