Posts tagged “world-models”

Your brain runs simulations before it acts. That is all a world model is and most of the systems worth simulating are things connected to other things, which is what a graph is for.

Instead of simulating the environment an agent acts on, simulate the execution itself — and find out whether a plan reaches the goal before running a single step of it.

The notation behind graph world models: the transition function, fixed versus dynamic edges, and where the training objective comes from.