[seminarios-mlpb] Tuesday, July 18th, 13:00, Anfiteatro do Complexo Interdisciplinar: Luis Sarmento on "Going Neurotic With Neural Word Embeddings... again!"

  • From: Pedro Balage <pedro.balage@xxxxxxxxxxx>
  • To: <seminarios-mlpb@xxxxxxxxxxxxx>
  • Date: Tue, 11 Jul 2017 16:24:31 +0100

Hi,

The next seminar will be given by Luís Sarmento on "Going Neurotic With Neural Word 
Embeddings... again!" (see information below).

For logistical reasons only (if you won't be able to do it you are welcome 
anyway), we ask you to register for the seminar in the link below, where you 
can inform us which type of food you would like to eat:

https://www.eventbrite.pt/e/going-neurotic-with-neural-word-embeddings-again-tickets-36107419263?aff=seminariosmlpb

Please, try to register up to this*Friday 11 am*, so that we can guarantee 
lunch for everyone.

Thanks,

Pedro Balage


********************************************************************
TALK ANNOUNCEMENT  --*Anfiteatro do Complexo Interdisciplinar*
********************************************************************

Priberam Machine Learning Lunch Seminar
Speaker: Luís Sarmento (Tonic App)
Venue: IST Alameda, Anfiteatro do Complexo Interdisciplinar
Date: Tuesday, July 18th, 2017
Time: 13:00
Lunch will be provided

*Title:*
Going Neurotic With Neural Word Embeddings... again!
*Abstract:*
Word embeddings, such as Word2Vec or Glove, are vector representations that 
capture lexical-semantic properties of words. They constitute a practical way 
for transferring knowledge between two machine learning models, and they 
contribute to greatly reducing the learning time required for solving various 
NLP tasks. There is great practical interest in experimenting with different 
word embedding models.
Neural-based models, due to their flexibility, are a great framework for  that 
experimentation. However, that very same flexibility also brings many degrees 
of freedom to the experimentation, which end up becoming a challenge in itself. 
In this talk, we will present Syntagma, a python toolkit (still under 
development) that enables rapid experimentation of neural word embedding 
models. We will present preliminary results of experimenting with some of the 
hyper-parameters of a baseline word embedding model (similar to Word2Vec), and 
we will discuss the next steps for Syntagma.

*Short Bio:*
Luis Sarmento holds a PhD in Computer Science from University of Porto (2010), 
with background in Electrical Engineering (Bs+MsC) and Artificial Intelligence 
(MsC). He has been working in the fields of Natural Language / Search for about 
15 years, both as a member of research groups at the University of Porto and in 
the industry. In 2010 he joined Portugal Telecom / SAPO as tech lead for 
Big-Data and Recommender Systems, and in 2012 he joined Amazon where, until 
early 2017, he lead research teams in the fields of Query Understanding and 
Voice Shopping. He in now CTO of Tonic App (http://www.tonicapp.com/), a 
startup developing productivity tools for medical doctors.
*Eventbrite:*
https://www.eventbrite.pt/e/going-neurotic-with-neural-word-embeddings-again-tickets-36107419263?aff=seminariosmlpb

********************************************************************
General info about our seminars:  
http://labs.priberam.com/Academia-Partnerships/Seminars
Please forward this link to whoever might be interested in attending the 
seminars.

Other related posts:

  • » [seminarios-mlpb] Tuesday, July 18th, 13:00, Anfiteatro do Complexo Interdisciplinar: Luis Sarmento on "Going Neurotic With Neural Word Embeddings... again!" - Pedro Balage