What's your project called?
Terrorism Entity Extractor
What is your chosen problem statement?
Subtheme 2: Strengthening Internal Security. Design and implement an LLM that can extract entities from reports into a knowledge graph, and an LLM that can answer questions based on the generated knowledge graph.
How does your hack answer the problem statement?
Our hack would be able to extract relevant entities from given articles or documents such as the name of person who attacked, the location where the terrorism attack occurred, the event that happened and the object used in the attack. Users would be able to interact with a chatbot to request information from an article given by the user, then the Named Entity Recognition (NER) model analyses and extracts the relevant entities from the article. Afterwards, the entities are deposited into a graph database along with their relationships with each other and the knowledge graph from the entities are given to the user. User would be able to ask subsequent questions such as "Who perpetrated the attack?".
How did you build your hack?
We built the hack using python and Google Colab. We implemented a pretrained NER model called Spacy and used it to extract relevant entities from articles. Then, we used Neo4j to create the graph database containing the entities and used Matplotlib to plot the knowledge graph. Afterwards, we designed a chatbot with Rasa to be able to handle user queries.
Log in or sign up for Devpost to join the conversation.