COVID update: Due to the pandemic the course will be online, keeping all planned activities. Instead of physical attendance, the classes will be broadcasted live online. The practical labs will be also held online, in two split groups with one lecturer in each. Given our online teaching experience these last weeks, we are confident that we will be able to offer a high-quality and engaging course, both at the theoretical and hands-on practical sessions.

Deep Learning neural network models have been successfully applied to natural language processing, and are now changing radically how we interact with machines (Siri, Amazon Alexa, Google Home, Skype translator, Google Translate, or the Google search engine). These models are able to infer a continuous representation for words and sentences, instead of using hand-engineered features as in other machine learning approaches. The seminar will introduce the main deep learning models used in natural language processing, allowing the attendees to gain hands-on understanding and implementation of them in Keras.

This course is a 20 hour introduction to the main deep learning models used in text processing, covering the latest developments, including Transformers and pre-trained (multilingual) language models like GPT, BERT and XLM. It combines theoretical and practical hands-on classes. Attendants will be able to understand and implement the models in Keras.

Student profile

Addressed to professionals, researchers and students who want to understand and apply deep learning techniques to text. The practical part requires basic programming experience, a university-level course in computer science and experience in Python. Basic math skills (algebra or pre-calculus) are also needed.

Contents

Introduction to machine learning and NLP with Keras

Machine learning, Deep learning
Natural Language Processing
A sample NLP task with ML
. Sentiment analysis
. Features
. Logistic Regression
LABORATORY: Sentiment analysis with logistic regression

Multilayer Perceptron and Word Embeddings

Multiple layers ~ Deep: MLP
Backpropagation and gradients
Learning rate
More regularization
Hyperparameters
Representation learning
Word embeddings
LABORATORY: Sentiment analysis with Multilayer Perceptron

Recurrent Neural Networks, Seq2seq, Neural Machine Translation

From words to sequences: RNNs
. Language Models (sentence encoders)
. Language Generation (sentence decoders)
. Sequence to sequence models and Neural Machine Translation (I)
Better RNNs: LSTM
LABORATORY: Sentiment analysis with LSTMs

Attention, Better Machine Translation and Natural Language Inference

Re-thinking seq2seq:
. Attention and memory
. State of the art NMT with Transformers
Natural Language Inference with siamese networks
LABORATORY: Attention Model for NLI

Convolutional Neural Networks. Pre-trained Language models.

CNNs for text
Pre-trained language models, BERTology.
Multilingual transfer learning
Deep learning frameworks
Last words
LABORATORY: Pre-trained transformers for sentiment analysis and NLI

Instructors

Person 1

Eneko Agirre

Professor, member of IXA

Person 2

Oier Lopez de la Calle

Postdoc researcher at IXA

Person 2

Ander Barrena

Postdoc researcher at IXA

Invited talk (Friday 14:30)

Unsupervised Machine Translation

While modern machine translation has relied on large parallel corpora, a recent line of work has managed to train machine translation systems in an unsupervised manner, relying on monolingual corpora alone. In this talk, I will introduce cross-lingual word embedding mappings, which are the basis of these systems, and present our work on both unsupervised neural and statistical machine translation.
Person 4

Mikel Artetxe

PhD student at the University of the Basque Country (UPV/EHU), Facebook Fellow

Practical details

General information

COVID update: Instead of physical attendance, the classes will be broadcasted live online. The practical labs will be also held online, in two split groups with one lecturer in each.

Bring your own laptop (in order to do the practical side).
Part of the Language Analysis and Processing master program.
5 theoretical sessions with interleaved labs (20 hours), plus an invited talk.
Scheduled from July 1st to 3rd 2020, 9:00-13:00 14:30-18:30 (Friday ends 16:00).

Where: "Ada Lovelace", Computer science faculty, San Sebastian
(practical classes will be held in labs, split groups).
The university provides some limited information about accommodation in San Sebastian (Basque/Spanish) and the Basque Country (English).
Lunch on your own in one of the cafeterias on campus.
Teaching language: English.
Capacity: 60 attendants (First-come first-served).
Cost: 274 euros (270 for UPV/EHU members).

Registration

Registration open: now to the 20th of June 2020 (or until room is full).
Please register by email to amaia.lorenzo@ehu.eus (subject "Registration to DL4NLP" and CC e.agirre@ehu.eus).
Same for any enquiry you might have.
The university provides official certificates. Please apply AFTER completing the course.
Public universities are not allowed to produce invoices, but we can provide a payment certificate.


Prerequisite
Basic programming experience, a university-level course in computer science and experience in Python. Basic math skills (algebra or pre-calculus) are also needed.
Bring your own laptop (no need to install anything).

Previous editions

Online class of July 2020 (left), with a handful of the 70 participants. To the right the screen lecturers had to talk to :-)

Class of January 2020

Class of July 2019