From UBC Wiki


Authors: Renee Lau, Jason Huang, Jerry Chan

What is the problem?

In a world where dating sites are prevalent, there are few such sites to help people find the perfect dog. Dogs come in different shapes, sizes, and personalities. Some are more trainable, others less so. Many require lots of exercise while a select few are quite happy to sit on the couch with their owners. This interface allows users to query for dog breeds which match the criteria specified, allowing them to find the most compatible companion in their lives.

What is the something extra?

Our program will generate the dog breeds that match the user's preferences (size, fur colour, energy level, etc.).
These results will be ranked by similarity to the given criteria.
Example Question: What is a breed that is medium sized and has big ears?

Also, added and modified the parsing of questions to create an interface where users can just type in the question.

Added extra ask forms to be able to process questions starting with "How does" and "Is".

What did we learn from doing this?

This project cemented our understanding of Natural Language Interfaces. It also helped us learn how to optimize the relations such that it was more extensible than the original interface shown to us.

In addition, we learned how to parse queries so that the user does not have to pass in a comma separated list with a predicate.

We learned how to add additional relations to process different types of questions.