Overview

Natural Language Processing is about techniques to automatically extract relevant information from language, primarily text. Recent advancements in NLP take the perspective of representation learning for language, where language is represented by learned vectors in a high dimensional space. In this workshop we will look at how these learned representations of language can be used to search for semantically similar pieces of text. We will look at baselines techniques such as TF-IDF and compare it to contemporary methods using pre-trained Transformers.

We will work in Jupyter Notebooks with standard python libraries for contemporary NLP such as GENSIM and Hugging Face Transformers. As an application we will look at patent data and how NLP can be used to create semantic representation of patent applications for similarity search.

Prerequisites

You should be fluent in python and familiar with numpy. The workshop will assume you have a basic grasp of linear algebra and probability theory.

Agenda

For updated agenda please visit https://enccs.se/events/2022-05-contemporary-nlp/

For questions regarding this event please contact us at [email protected].

This training is intended for users established in the European Union or a country associated with Horizon 2020. You can read more about the countries associated with Horizon2020 here https://ec.europa.eu/info/research-and-innovation/statistics/framework-programme-facts-and-figures/horizon-2020-country-profiles_en

 
https://events.prace-ri.eu/event/1382/