Some code for a SIMPLIFIED version of the Harry Potter Sorting hat #STEM project that leverages IBM Watson services
Back in 2016 - IBM made available a pretty innovative (for the time) natural language classifier - but allowed for optimization and training (weights) for certain topics and idea ideas; by uploading a CSV file to the existing corpus of the classify - the gen-1 AI / ML could be trained to do inference and provide outputs as desired; the knowledge graph at the time was quite powerful - state of the art for 2016 2017
We thought it would be fun to try to do a classification for the Harry Potter sorting hat, which is out of the movie - which briefly put - read the minds of the Hogwarts students - and classifies them in the correct house; in this case, we were going to use the utterances by speech of the wear of the hat substitute for the mind reading;
The chassis hardware and pieces were prototyped by my daughters in our home lab - the ground truth was written by my eight-year-old daughter (a harry potter SME) and performed very well.
I brought the project prototype to science hack day where the team really took it to the next level!
From there, there were some offshoot projects in the open-source community and with Ibm;
Nearly 10 years later with the advent of large language, models, and next generation AI - we've come back to take a look at this project - as it provides a fun foundation for experimentation
"how to" video series: https://www.youtube.com/watch?v=6mV4qbgKYtk
original test from science hack day: https://www.youtube.com/watch?v=ns3fn_2ziUE
Here's the CSV file (ground truth) that is used to train the IBM Watson NAtural LAnguage Classifier (you can modify) https://github.com/rustyoldrake/Harry_Potter_Sorting_Hat_Simple/blob/master/sorting_hat_NLC_ground_truth.csv
Natural Langauge Classifier (NLC) SERVICE (IBM Watson API) (( Decommisioned )) https://console.bluemix.net/docs/services/natural-language-classifier/getting-started.html#natural-language-classifier
There was a team of about a half a dozen people who were part of our team at Science Hack Day who played an instrumental role in getting Python bits working (Thank you!)
=== Natural Language Classifier: https://ibm.box.com/s/rdlog2sue79178816s0rabkbi7ifu5vg - Handbook https://github.com/rustyoldrake/NLC_ground_truth_wonderland http://emojiburger.mybluemix.net/ https://dreamtolearn.com/ryan/data_analytics_viz/98 https://dreamtolearn.com/ryan/r_journey_to_watson/13 (ctrl-f for ’NLC)
"The Sorting Hat does everything in the Harry Potter movie ought to, except read the wearer’s mind. By integrating IBM Watson, Natural Language Classifier (NCL), Raspberry Pi 2B+ project board, Adruinos, YUN, servos, LEDs, microphones, speaker, batteries and speech-to-text/text-to-speech services, the wearer is able to express characteristics of him or herself and gets sorted into the most appropriate house he or she belongs to."
