Cluster 12: Machine Learning: Can We Teach a Computer to Think?
Week 2
Friday (7/14)
Cluster 12 started their Spotify Project today which is an open-ended project that allows them to play around with the Machine Learning models they learned such as k-NN, k-Means Clustering, Decision Trees, Neural Networks, and Logistic Regression. Students started to play around and ran into issues with the spotipy library, but with the help of the TA’s they were able to fix it. After the lab, Cluster 12 stole the show at COSMOlympics with their skit about three COSMOS runaways. They even got their teaching assistants Sarayu, Emmett, and Owen in the skit and Teacher Fellow Shirley's daughter to play the train whistle!
Monday (7/17)
Students started their morning with learning about Logistic Regression, a classification model to predict the probability of an event given a dataset. Students spent the rest of class working on their Spotify projects with their partners and the Ethics Essay due on Friday. The projects consisted of a variety of topics such as a Genre-Generator and a Spot-a-Friend applying the ML models students learned. As fun as the project seems, many ran into an intense amount of error messages from the code, but with the help of the TA’s and Professors, students were able to overcome the challenges and debunk the code.
Tuesday (7/18)
Cluster 12 was excited for another lecture with Professor Smarr, an Assistant Professor in the Bioengineering and Data Science department at UCSD. He discussed the Circadian Rhythm, a natural behavioral and physical cycle that occurs in living things. After the lecture, students had more time to work on their Ethics Essay and Spotify project. Right after lunch, students went on their cluster field trip to Google and got to explore the building. They then had the opportunity to listen to a presentation on Machine Learning and its applications and got their questions answered in a Q&A session.
Wednesday (7/19)
In the morning Professor Curt gave a lecture discussing machine learning using photos, including how photos and colors are expressed on computers and experimenting with Google Quickdraw. Then, time was given to work on the ethics essay and wrap up the final touches on their Spotify projects. Presentations took a little longer than expected due to many going over the time limit. There were many interesting projects that ranged from recommended you a song based on your music taste, predicting what scene a video game was in when a song was playing, and code that paired two people together based on music taste. After lunch, students started their second project using image classification.
Thursday (7/20)
Cluster 12 began the day by listening to a Zoom panel with COSMOS alumni. They discussed their experiences at COSMOS and shared advice. Awards for the Spotify project were announced after. Congratulations to Angie and Isaac for winning the People’s Choice Award and to Faith and Ronit for winning the Faculty Choice Award! Then, students worked on their ethics essay for the last time in class. After lunch, students continued work on their image classification projects.