Intro to deep learning for question answering

This is a past event

85 people went

Universitatea Politehnica din Bucuresti, Facultatea de Automatica si Calculatoare

313 Splaiul Independentei · Bucharest

How to find us

Facultatea de Automatica si Calculatoare se afla la intrarea dinspre Bd. Iuliu Maniu a UPB. Ne vom intalni in cladirea noua, aflata chiar langa bulevard, sala PR002 (amfiteatrul mic de la parter)

Location image of event venue


The talk will provide the basics for understanding current question answering solutions using deep learning (convolutional, recurrent NNs, with and without attention, etc.)

Although this talk is introductory, presenting papers from about 2 years ago, it will allow the participants to understand more recent advancements from 2016 which will be presented at a subsequent Meetup in Feb/Mar 2017.

For the current Meetup, we propose three different solutions:

- a simple CNN for text [1]

- a dependency tree - RNN [2]

- a LSTM-based solution [3]

Notice that this meetup is taking place at University Politehnica of Bucharest, Faculty of Automatic Control and Computer Science, PRECIS (new) building, room PR001 (large amphitheater on the ground floor)


Suggested reading materials:

[1] Yu, Lei, Karl Moritz Hermann, Phil Blunsom, and Stephen Pulman. "Deep learning for answer sentence selection." arXiv preprint arXiv:[masked] (2014).- online here:

[2] - Iyyer, Mohit, Jordan L. Boyd-Graber, Leonardo Max Batista Claudino, Richard Socher, and Hal Daumé III. "A Neural Network for Factoid Question Answering over Paragraphs." In EMNLP, pp. [masked]. 2014 - online here:

[3] - Tan, Ming, Bing Xiang, and Bowen Zhou. "LSTM-based Deep Learning Models for non-factoid answer selection." arXiv preprint arXiv:[masked] (2015) - online here:

Speaker: Traian Rebedea (+ Stefan Ruseti)