Weight-Space World Models
Weight-Space World Models
Auto-generated stub. Edit this file to add more details.
A world modeling framework that encodes environment states into the weights and biases of coordinate-based implicit neural representations.
Why It Matters
This introduces a paradigm shift in world modeling by encoding environment states directly into the parameters of a coordinate-based neural network rather than using traditional latent vector spaces, effectively bypassing the decoder bottleneck.
Evidence
represents the system state as the weights and biases of an auxiliary coordinate-based implicit neural representation (INR)
Related Papers
Metadata & Links
- created_at
- 2026-05-10T06:06:03Z