This week in AI & Machine Learning: The NFL tackles CTE with AI, A roomba that avoids dog poop, NOLS keyhole imaging, AI that writes its own code, deep reinforcement learning, and more!
My Top AI Highlight:
NFL and AWS Launch Artificial Intelligence Challenge to Crowdsource Ways to Automate Player Identification using NFL Game Footage
The NFL kicked off this week, and in partnership with Amazon Web Services have launched an open source contest that aims to help reduce injuries on the field–specifically identifying and tracking players involved in helmet impact.
The initiative aims to help the NFL monitor and prevent occurences of Chronic Traumatic Encephalitis or CTE, a so far incurable and minimally treatable, brain injury known to occur in people who experience repeated head trauma.
Simulation of a Computer Vision Model analyzing game footage.
” This competition is foundational in helping identify each player’s risk to injury-causing events, especially when it comes to head health.”
The contest is open until November 2nd and there is $100,000 in total prize money up for grabs. You can check out the official rules here.
🤖 Artificial Intelligence News:
- iRobot’s newest Roomba uses AI to avoid dog poop
- NOLS keyhole imaging can see inside a closed room
- Keurig’s BrewID makes coffee exactly as it was intended
- OpenAI’s Codex: AI That Writes Its Own Code
- IBM Watson Helps ESPN Fantasy Football Players make Better Trades
🛠️ Developer Tools & Education:
- Personalized ASR Models from a Large and Diverse Disordered Speech Dataset
- Deep Reinforcement Learning With Jetbot
🎤 Interesting Podcasts & Interviews:
- Data Science Podcast with Pat Walters (Cheminformatics Scientist)
- Deep Reinforcement Learning for Game Testing at EA with Konrad Tollmar
- Asking A Lawyer About GitHub Copilot
- Deep Reinforcement Learning for Robots
📄 Notable Research Papers:
- SONIC: A Sparse Neural Network Inference Accelerator with Silicon Photonics for Energy-Efficient Deep Learning
- Learning from Uneven Training Data: Unlabeled, Single Label, and Multiple Labels
- NEAT: Neural Attention Fields for End-to-End Autonomous Driving
- ConvMLP: Hierarchical Convolutional MLPs for Vision
