Big thanks to Social Tables for hosting!
Food and drinks are sponsored by the pseudo-anonymous Jay Kay! Thanks Jay Kay!
Want to be an official sponsor? Check out https://dc.hackandtell.org/sponsor.html
Event link: https://www.meetup.com/DC-Hack-and-Tell/events/264249560/
Notes copied directly from the hackpad!
Jessica Wherry
Ain’t Nobody Got Time for That
There’s a lot of language in terms & conditions. No one reads it all. You need two months to read through all ToS you agreed in a year.
Kit Walsh at EFF pays attention to things. Examples:
- Facebook can use your name and likeness in Ads
- AWS is not responsible for any data loss!
- OnStar can sell your location to anyone
- https://tosdr.org/
UsablePrivacy.org - annotated rules
https://www.youtube.com/watch?v=ZcjtEKNP05c
Q: is there a corpus of previous ELUAs and TOS for companies? For instance, How has PayPal’s terms changed/grown over the years?
Rachel Sweeney
https://github.com/rwsweeney/face_recognition_greeter
I’ve taken an old android phone, and have trained a machine learning model to recognize my wife and I. I’ve also created a few different functions that make api calls to wikipedia to pull down lists of “Sarah” and “Rachel”, randomly choose one and then output the wikipedia summary. Finally I’ve passed that output to aws polly to synthesize the text to speech. The plan is to point it at our front door to tell us fun stuff when we get home, but the first model I trained wasn’t as good as I would like. About to start retraining another model!
Amazon Polly - handy text to speech web service. Free for the first 4 million characters per month.
Zuri Hunter
I would like to share the trials and tribulations that I encountered in building my side project “Curl Critic”. Curly is a hair pattern. Lot’s of extra steps needed to take care of curly hair. Sulfates bad! Other chemicals??? Who knows, have to be careful…
Rating system for hair products, esp for curly hair. Takes information from Instagram and Twitter to rate products. API that takes data from social media, processes it using sentiment analysis. Lot’s of work needed for the data, NLP takes care. Raw twitter data is messy. Need to find cleaner data sources or find a better way to get a sense of what’s being said.
Byron Gaskin
I’ve been playing around with the StyleGAN code to generate hyperrealistic fake faces. I’ll present some of the cool/weird things I found while trying to make a deepfake of myself.
Travis Hoppe
BEFRAME is an AI-powered video project to clip ONLY the scenes in the movie where the character is framed and not speaking. They can just “be” in the frame. Explore the characters and director’s choices from Legally Blonde, The Exorcist, Fight Club, Pitch Perfect, Die Hard, Pretty Woman, The Princess Bride, and Requiem For a Dream.
Each movie is first clipped by visual content, and then analyzed for shot type. Only the Medium Close-up (MCU) shots are preserved. Google’s speech detection is used to filter out any shots with detected words. Finally, the shots are strung back together in sequence.
video: https://www.youtube.com/watch?v=K0_O34eoC68&feature=youtu.be