Atlas is demonstrating reinforcement learning policies developed using a motion capture suit. This demonstration was developed in partnership with Boston Dynamics and RAI Institute.

Meet Penelope, the dielectric kitty. One of our teams - Emily, Anqi, Lauren, and Margaret - are developing a long range (kms) cat tracker that utilizes LoRa. Emily needed to test the antenna performance vs orientation on a kitten-like dielectric equivalent. So meet Penelope!

The best time of year, all the projects are coming alive. LQR, system ID, Deep RL… all in a little unicycle running on an @NVIDIA Orin. Lots more to share on this one by ENPH 479 students Jackson Fraser, Julian Lapenna, Kyle Mackenzie, Simon Ghyselincks, and Tristan Lee.

And here we have the neuromorphic compute team benchmarking an @NVIDIA Jetson Orin against their custom designed FPGA based neural network inference system running on AMD Versal AI Edge Gen 2 system. The driving environment is running on Gazebo Simulator and ROS (Robot Operating System (ROS)).

Beer Pong Sentinel! Polina, Eldad, Harry, and Brian are using two 500 frames per second Teledyne FLIR cameras for stereoscopic vision (thank you opencv), high power motors and a BB gun to create an air interdiction region in which no ping pong ball can enter.

A bucolic scene: dog owner with dog circa 2025. 300nm to 800nm photons on CCD sensor. Artist unknown. Unitree Go2 trained to walk using deep RL.

Rich's slogans for AI research (revised 2006): 1. Approximate the solution, not the problem (no special cases) 2. Drive from the problem 3. Take the agent’s point of view 4. Don’t ask the agent to achieve what it can’t measure 5. Don't ask the agent to know what it can't verify

Awesome work. Great to see Dreamer in Nature! Congrats Danijar Hafner Timothy Lillicrap et al. Nature news article on it with quotes from yours truly. Here is the quote I provided that did not make it into the article: I love this work. In fact, the Dreamer work is some of my favorite

We caught up with Joel T. testing his ghost buster system outside the Project Lab. Joel and 3 other classmates designed their own unexploded ordnance (#uxo) detection system for their capstone project - which might or might not include a diffusion based deep learning model.

🚨 New: We a16z built an 8x RTX 4090 GPU AI workstation from scratch —compatible with the new RTX 5090 with PCIe 5.0, for training, deploying, and running AI models locally— so you don’t have to. Here’s how we built it, why it matters, and how you can build one too. Full

badminton robot from ETH Zurich