Products · P.01 · In research
Causal State Delta
Causal State Delta is an approach, in research at Arxigen, for AI agents that operate games, websites and 3D software. Instead of re-analysing the whole screen every frame, the agent reads only the changes that matter to its next move.
The problem
An agent that acts on a screen usually looks at every frame in full, even when very little has changed. Much of that work has no bearing on what the agent should do next.
The approach
Causal State Delta focuses the agent on the parts of the screen that changed and that matter to its next action, and leaves the rest alone. The aim is faster decisions at the same accuracy.
Who it is for
Teams building AI agents that operate games, websites and 3D software.
Where it comes from
It grows out of Arxigen’s research in Efficient AI: inference acceleration, optimisation and model efficiency.
Status
In research. There are no published results yet, and figures shown in the site’s animations are illustrative.
Early access
Ask for early access, or propose a research collaboration, through the contact form. Arxigen will be in touch when Causal State Delta is ready.