Hack Cambridge 2017 was a huge success. You can see the list of submissions at https://recurse.devpost.com/submissions. The inventiveness behind many of these ideas is impressive as is the amount of code that was generated in 24 hours. Don’t let anyone say that the exciting story of “computing meets world”, that began more than 30 years ago, has run its course. Far from it.
We gave a short talk “How to Create an Audio Sentence Finisher using Machine Learning” and demonstrated this with Maybot – a neural network trained on Theresa May’s spoken contributions in Parliament scraped from Hansard. We used the network in RoboTrumpDNN https://github.com/ppramesi/RoboTrumpDNN as a starting point for this. You can read the presentation here. If you are thinking of creating something similar, the links on the last page of the presentation should steer you in the right direction.
Our mentors helped a number of teams and there were some good examples of machine learning.
We’d particularly like to congratulate Peter Carragher, Ramona Comanescu, Andreea Cucu and Arthur Wilcke from the University of Edinburgh for their project “Cyber Mum”. They picked up our challenge of doing something interesting with machine learning and took our Maybot code and made an Alexa skill that creates interactive fairy tales. Their submission page is at https://devpost.com/software/cyber-mum. The code is available at https://github.com/PeterCarragher/HackCambridge.
See the video (courtesy of Syeef Karim of Major League Hacking)
See video on King James (the predecessor to Maybot)
Congratulations to the Cyber Mum team for winning the prize for the “Best Use of Amazon Web Services”.