Automated Content Understanding

Going from a normal website to using a chatbot to access the same information. This is a process we often go through. In this project you will try the latest AI methods to (partially) automate this process.

Often a chatbot is placed on a website to help people find information which is also available elsewhere on the website. However, training the chatbot will require you to input this information, which is mostly a manual task. The goal of this project is to automate this process. First we crawl information that is available on a website, think about FAQs, help sections, and pages with information. This information needs to be cleaned and translated into a format the chatbot understands. Extracting information can be done with different natural language processing techniques. After the content is understood, it needs to be transformed to user-intents that are used to train the chatbot. Here we need paraphrasing and sentence generation to generate questions for the chatbot to answer. It is also possible to do a research project for one of the subjects.

For this project, experience with one of the following is recommended:

Web scraping
Text classification / data mining
Back-end frameworks (Node.js, Python, Java, etc.)
Familiarity with Cognitive Services (Microsoft, IBM)

Are you curious about this assignment, or one of our other assignments? Do you have ideas for your own project related to AI, IoT, VR/AR, Block-chain, Mobile, Web, or some other new technology, just send us a mail, or stop by at our office. We will try to help you realize your ideas!