CLS Spekaer Series: Angeliki Lazaridou, February 12, 4pm

[Raquel Fernandez]

Dear all,

We are very happy to announce that Angeliki Lazaridou from Google
DeepMind will give a talk at the ILLC Computational Linguistics Seminar
next week. You can find the details below.

http://projects.illc.uva.nl/LaCo/CLS/

Angeliki Lazaridou, Google Deep Mind
12th February, 2019, 4pm, room F3.20, Science Park 107

*Emergence of (linguistic) communication through multi-agent interactions*

Distributional models and other supervised models of language focus on
the structure of language and are an excellent way to learn general
statistical associations between sequences of symbols. However, they do
not capture the functional aspects of communication, i.e., that humans
have intentions and use words to coordinate with others and make things
happen in the real world. In this talk, I will present two studies on
multi-agent emergent communication, where agents exist in some grounded
environment and have to communicate about objects and their properties.
This process requires the negotiation of linguistic meaning in this
pragmatic context of achieving their goal. In the first study, I will
present experiments in which agents learn to form a common ground that
allow them to communicate about disentangled (i.e., feature norm) and
entangled (i.e., raw pixels) input. In the second study, I will talk
about properties of linguistic communication as arising in the context
of self-interested agents.

--
Raquel Fernandez
Institute for Logic, Language & Computation
University of Amsterdam
http://www.illc.uva.nl/~raquel

[Raquel Fernandez]

Dear all,

A reminder that Angeliki Lazaridou from Google DeepMind is our invited
speaker today at the ILLC Computational Linguistics Seminar. You can
find the details below. You are also welcome to join for drinks at
Polder from 6 to 7 pm.