It’s free, every week, in your inbox.
And, outside of simulations, anything can happen.
Agents have to deal with gravity, unexpectedly slippery surfaces, and unplanned interference from other agents.
As the AI starts its training, its barely able to move its physics-based humanoid avatar around the field.
Our work constitutes a complete demonstration of learned integrated decision-making at multiple scales in a multiagent setting.
Quick take:This work is pretty rad.
But were not so sure it represents a complete demonstration of anything.
The model is obviously capable of operating an embodied agent.
The bottom line here, is that the AI isnt learning how to play football.
Its brute-forcing movement within the boundaries of its simulation.
I dont know what its running away from, but Im certain that its the scariest thing ever.
However, its almost certain that these models will become more robust as time goes on.
Weve seen what Boston Dynamics can do with machine learning algorithms and pre-programmed choreography.
