To celebrate the 25th anniversary of the Auton Lab, I organized the lab's first-ever, data science hack-a-thon, otherwise known as the hackAuton, focused on AI with real world impact. Participants broke into teams on Friday night and had until Sunday afternoon to work on one of seven challenge problems.
We received generous donations from multiple financial sponsors. We invited all sponsors to send a representative to sit on the judging panel and to submit a dataset and challenge problem for consideration.
Making sure everyone was well-fed, had fun, and learned something was our priority. Therefore, we focused on creating a celebration (positive experience) rather than a competitive prize.
I differentiated our hackAuton from other similar events in that the focus was making sure each team had something to show for their efforts. We set up a GitHub page which hosts the submitted work from each team, and some of the teams even published their work in a writeup on arXiv. We also recorded the presentations (need a Box account to view) from each team and provided links to them online. Having all these pointers to participants' hard work made it possible for them to share their achievements with others on their resume.
Thankfully, we partnered with our friends at the Pittsburgh Supercomputing Center to provide cloud computing credits for all the event participants. Each team had access to as many CPUs and GPUs that they could handle. Normally, this would have been a major financial consideration for the event, so we are especially grateful to PSC for their resources and their support for all using their servers.
One of our sponsors, Connectas, is a news organization focusing on covering developments in the Americas. In addition to submitting a challenge problem for the event, they even wrote up a summary of the event.
For more details about last years event, you can check out the hackAuton website: 2018.hackAuton. We are planning on running our second hackAuton near the start of the fall semester.
There were two best-in-show winning teams, Deep Girl Network and Team X. Both of their presentations are linked below:
Deep Girl Network
Team X
