The course will be offered online, both theory sessions and practical labs. We will try offer an engaging course, both at the theoretical and hands-on practical sessions.

Language Technology is increasingly present in many of the applications we use in our everyday activities (Google Home, Amazon Alexa, Siri, Google Translate, Grammar checkers, Google search engine, ChatGPT...) and the need of experts that can develop applications based on Language Technology is an ever growing demand both in the industry and academia. This course will introduce the most commonly used techniques to build applications based on Language Technology. Thus, the attendees will learn how to apply techniques such as document classification, sequence labeling, as well as vector-based word representations (embeddings) and pretrained language models for core applications such as Opinion Mining, Named Entity Recognition, Fake News Detection, Question Answering or Text Generation..

The course will have a practical focus (laboratories and practical tasks) learning to use readily available LT toolkits (Spacy, Flair, HuggingFace Transformers) based on machine and deep learning in a multilingual and multi-domain setting. The aim is to allow attendees to acquire the required autonomy to solve practical problems by applying and developing Language Technology applications. The course will be taught in English.

The course is part of the NLP master hosted by the Ixa NLP research group at the HiTZ research center of the University of the Basque Country (UPV/EHU).

Student profile

This course is targeted to graduate students and professionals from a range of disciplines (linguistics, journalism, computer science, sociology, etc.) that need an applied introduction to Language Technology. This involves identifying the required linguistic resources, appropriate tools/libraries and techniques with the aim of acquiring the required autonomy to solve practical problems by applying and developing applications based on Language Technology in different and creative ways.

For the practical content (coding exercises) some experience in python programming is recommended. Previous attendance to the Deep Learning for Natural Language Processing course is might be useful although not required.

Contents

Introduction to Applications of Language Technology

Natural Language Processing
Cross-lingual Information Extraction
LABORATORY: Stance detection with logistic regression
. Features
. Static Word Embeddings
Introduction to Flair
Introduction to Spacy

Text Classification

Fake News, Stance and Propaganda
Detection
. Fake News
. Hyperpartisanism
. Hate speech
Inference
. Fact-checking
. Stance
. Argumentation
LABORATORY: Stance Detection
. Training with Flair and Spacy

Sequence Labelling

Named Entity Recognition
. Contextual Word Representations
. Datasets
. Evaluation
Morphology
. Contextual and neural lemmatization
. Evaluation and application to high-inflected languages
LABORATORY: Train language independent neural sequence taggers with Flair and Transformers
. Named Entity Recognition
. Contextual lemmatization.

Opinion Mining

Fine-grained Sentiment Analysis
Aspect-based Sentiment Analysis
Multidomain and multilingual issues
LABORATORY:
Sentiment Analysis
. Text Classification
Opinion Targets and Aspects
. Sequence Labelling with Transformers

Question Answering

Redefining NLP tasks as QA
Pre-trained language models, Transformers
Multilingual transfer learning
Last words
LABORATORY Build and train a Question Answering system with encoders such as BERT.

Text Generation

Tasks based on Text Generation
Pre-trained language models, encoder-decoder (T5) and decoder (GPT) models
Multilingual transfer learning
Last words
LABORATORY: Automatic Argument generation using mT5.

Instructors

Person 1

Rodrigo Agerri

Permanent researcher, member of Ixa
and HiTZ

Person 3

Irune Zubiaga

FPI researcher, member of Ixa
and HiTZ

Practical details

General information

The classes will be held online. The practical labs will also be online.

Part of the Language Analysis and Processing master program.
5 theoretical sessions with corresponding programming labs (20 hours).
September 16th to 20th 2024, every afternoon from 15:00 to 19:00 CET.

Course language: English.
Capacity: 60 attendants (First-come first-served).
Cost: 270 euros + 4 insurance (270 for UPV/EHU members or if you also apply to the other Specialization courses.).

Registration

Registration is open.
  • Please register by email to ixa.administratzailea@ehu.eus (subject "Registration to ILTAPP" and CC olatz.arregi@ehu.eus).
  • Also for any enquiry you might have.
  • After you receive the payment instructions you will have three days to formalize the payment.
  • The university provides official certificates (for an additional 27.96 euros). Please apply AFTER completing the course.
  • UPV/EHU can provide invoices addressed to universities or companies. More details are provided after registration is made.



Prerequisites
Basic Python programming experience.
Not a requirement but, previous attendance to the Deep Learning for Natural Language Processing course held the previous week will help students to better understand the underlying algorithms of Language Technology applications.
Bring your own laptop (no need to install anything).
Although not strictly necessary we recommend subscribing to Google Collab Pro for more GPU allowance.

Previous editions

Class of July 2022.

Class of July 2021.