Google DeepMind’s IQ test for generative AI models is interesting to say the least, as this new benchmark was designed specifically to assess their physics understanding. Called Physics-IQ, the dataset can only be solved by acquiring a deep understanding of various physical principles, like fluid dynamics, optics, solid mechanics, magnetism and thermodynamics.
Current models like Sora, Runway, Pika, Lumiere, Stable Video Diffusion, and VideoPoet, all have a limited understanding of physics, but some test cases can already be successfully solved. One example would be that fluid dynamics tend to work better than solid mechanics, indicating that acquiring certain physical principles from observation alone may be possible, but significant challenges remain. You can check out the Github page here.
- SIZE DOWN. POWER UP — The far mightier, way tinier Mac mini desktop computer is five by five inches of pure power. Built for Apple Intelligence.*...
- LOOKS SMALL. LIVES LARGE — At just five by five inches, Mac mini is designed to fit perfectly next to a monitor and is easy to place just about...
- CONVENIENT CONNECTIONS — Get connected with Thunderbolt, HDMI, and Gigabit Ethernet ports on the back and, for the first time, front-facing USB-C...
