$$Events$$

Oct. 29, 2019
13:00
-14:00

96-001

Speaker: Hila Gonen, Bar-Ilan University

Title:  Representations and Bias in NLP

Abstract:

Continuous text representation is a core-element across most NLP applications nowadays, especially since the reintroduction of Deep Learning (neural networks) models. In this talk I will focus on two aspects of representations in NLP. The first one is creating improved representations using representation learning. I will demonstrate that using multitask learning with multilingual corpora to improve preposition sense disambiguation in text. In the second part I will examine standard representations that are used in NLP, present biases they exhibit and discuss challenges in current debiasing methods. These two works together provide an interesting view on the challenge of creating useful representations, while being aware of social fairness aspects.

About the speaker:

Hila is a PhD student at Bar Ilan University in the field of Natural Language Processing and Deep Learning, under the supervision of Dr. Yoav Goldberg.
Prior to that, she obtained her M.Sc in Computer Science from the Hebrew University, under the supervision of Prof. Orna Kupferman.
She is fascinated by languages and interested in relations between different languages and the way multilingual signals can be used for various tasks. Lately, she has been focusing also on research in the area of gender bias in NLP.