An AI-Driven Chatbot for Refugees Helplines

Challenge Completed!

In this 8-week challenge, a global team of 50 AI engineers developed a chatbot to support refugees helpline.

The problem

The integration process for immigrants in a new country is not an easy task. Even with hard work and the desire to fit in, immigrants may face many challenges, including accessing basic information or understanding some of their basic rights in the country of residence.

With the help of Omdena Collaborators, Art For Impact wants to take a step further and utilize cutting-edge technology in building a bridge of understanding between cultures, combating hatred, showing the world that immigrants are contributing to their host communities, and allowing them to prosper. Currently, most of the conversations are being carried out by human agents and others by simple multiple option lists. The majority of the questions are repetitive. The goal is to reduce the time agents spend on simple and repetitive questions by building an intelligent Chatbot, able to hold intuitive and natural conversions to provide the services and answer those questions.

The project outcomes

Starting with data collection and augmentation, the team applied exploratory data analysis (EDA) to gain first insights into the data. Various EDA approaches (word clouds, topic modeling, knowledge graphs, etc.) were helpful in identifying relevant chatbot intents, entities, and responses.

Next, the team built NLU (natural language understanding) and NLG (natural language generation) models in Rasa. In order to handle emergency cases, the models identify utterances and messages related to emergencies and group them by intent. Keywords are highlighted in emergency messages as entities. Some stories were developed as conversation archetypes

Finally, the application was deployed using Docker.