Automatic facts extraction

The system extracts from text entities of different types (artist names, style of artist, locations and dates) and relations between them.

Problem

The objective of this project was the development of a system which helps to automatically extract structured information from raw text.

Solution

We applied our knowledge in NLP (Natural Language Processing) and developed neural network which processes input texts and classifies every input word (more formally, a NER system). We used unsupervised techniques (like Word2Vec) because we had small training dataset.