Inspiration
The whole idea came to us when our teacher who is preparing us to start in Olimpiad in Informatics had started to throw chalk in people who joke, say extraordinarly stupid things. So why couldn't robot do something like that? Now, it can (thanks to us).
What it does
It is based on algorithm which calculates something like probability of given sentence being nonsense. Algorithm gets input comprising of sentence in English. Then our algorithm decides if the sentence makes sense or not. As dictionary we used a Wikipedia wordbank with categories. To search for categories that are connected with categories of any given word we use some kind of DFS algorithm.
Then we put our algorithm into a kawasaki robot. Robot is behaving like our teacher, but not in 100%:
it awards you if you say something not stupid (by giving you sweets) and punishes you otherwise (by not giving you sweets).
How we built it
Using our best ideas we created a plan, which allowed us to succeed. Every person in our team was responsible for one task. We were divided into two subgroups: theoretical group and practical group. The first of them was responsible for algorithms, math background and optimalisation, the second for hardware, robots kawasaki, low-level programming.
Challenges we ran into
One of the most challenging problems that was on our way to serve our purpose was difficulties connected with natural language processing. Also calculating probability wasn't as easy as we thought because it depends on lots of variables. At the beginning we had hoped to program kawasaki robot to throw ball in people, but the idea turned out to be impossible because of robot's technical limitations, so we decided to use awards and punishments system.
Accomplishments that we are proud of
1) We had fun - it is the reason why we had started working on rohatyniec zaglady 3000 - we wanted to have some fun. 2) We have learned a lot and earned a lot of experience - it can help us in the future. 3) We are proud that our first serious project was complished (in demo version) on our first serious IT event.
What we learned
1) Some algorithms used in Natural Language Processing 2) Using Kawasaki robots 3) Using multiple programming technologies in one project (Python, C#, C ++, AS) 4) Team cooperation 5) Time managment 6) Using Wikipedia API
What's next for Rohatyniec zagłady 3000
So Rohatyniec zaglady 3000 is a first step to more practical usages like teaching people. (Such robot would understand what you talk to it and it would be able to determine if your speech is valuable). We are hoping to continue developing it in the future, especially improve nosence detection.
Log in or sign up for Devpost to join the conversation.