@causalwizard.bsky.social
In addition, by analyzing the divergence of the latent state we found evidence that the resulting plans (paths) are more consistent over time.
October 29, 2025 at 7:35 AM
In addition, by analyzing the divergence of the latent state we found evidence that the resulting plans (paths) are more consistent over time.
We found that the HRM-Agent can learn to navigate in dynamic and uncertain maze environments, with doors which open and close randomly.
October 29, 2025 at 7:31 AM
We found that the HRM-Agent can learn to navigate in dynamic and uncertain maze environments, with doors which open and close randomly.
The Hierarchical Reasoning Model (HRM) has impressive reasoning abilities given its small size, but has only been applied to supervised, static, fully-observable problems.
We wanted to see if we could train a HRM to navigate in a maze using only reinforcement learning.
We wanted to see if we could train a HRM to navigate in a maze using only reinforcement learning.
October 29, 2025 at 7:30 AM
The Hierarchical Reasoning Model (HRM) has impressive reasoning abilities given its small size, but has only been applied to supervised, static, fully-observable problems.
We wanted to see if we could train a HRM to navigate in a maze using only reinforcement learning.
We wanted to see if we could train a HRM to navigate in a maze using only reinforcement learning.