Projects
Politiquices.PT
Thousands of archived news titles were analysed to identify those reporting a supportive or opposing relationship between political entities and linking them to Wikidata, resulting in semantic SPARQL graph. The project was awarded the 2nd place in the “Arquivo.pt Awards 2021”.
nervaluate
A Python package for evaluating Named Entity Recognition (NER) models, as defined by the SemEval 2013 - 9.1 task. It goes beyond simple token-level accuracy, scoring predictions across multiple matching scenarios like: exact, partial, entity-type, and strict; giving a more realistic picture of NER model performance.
BREDS
An approach to extract named‑entity relationships without labelled data by relying on an initial set of seeds, i.e. pairs of named entities representing the relationship type to be extracted. The algorithm uses the seeds to learn extraction patterns and expands the initial set of seeds using distributional semantics to generalise the relationship while limiting the semantic drift.
snowball-extractor
Python implementation of Snowball, a bootstrapping/semi-supervised system for extracting relationships between named entities from large plain-text collections. It starts from a small set of seed entity pairs (e.g., company–headquarters) and iteratively learns extraction patterns to find new relationship instances in text.
ML Report Kit
A Python package for generating evaluation reports for supervised machine learning classifiers in just two lines of code. It produces classification reports (precision, recall, F1), precision-recall curves, confusion matrices, and per-sample prediction scores saved as images and CSV files for easy inspection and sharing.
Datasets
I maintain and curate public datasets and lexicons for NLP research:
-
Annotated Semantic Relationships Datasets - Traditional, Open, and Distantly Supervised relationship extraction datasets
-
NER Datasets - Named-Entity Recognition datasets organized by language: Portuguese, German, Dutch, French, English, Italian, Spanish, Polish and Russian
-
Lexicons - Portuguese person names, multilingual names, noun quantifiers, acronyms, and stop-words