Hitting Time Isomorphism for Multi-Stage Planning with Foundation Policies

Magnus Victor Boock Abdullah Akgül Mustafa Mert Çelikok Melih Kandemir · 2026

Read Hitting Time Isomorphism for Multi-Stage Planning with Foundation Policies by Magnus Victor Boock Abdullah Akgül Mustafa Mert Çelikok Melih Kandemir (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-06470.txt.

Description

Magnus Victor Boock, Abdullah Akgül, Mustafa Mert Çelikok, and Melih Kandemir learn codes for offline reinforcement learning from hitting times, the expected steps needed to first reach a state. Earlier geometry methods often produce symmetric distances or break the triangle inequality. Their framework instead builds a directed temporal geometry inside a Hilbert space.

They prove the code exists under a closure condition and is unique up to a bounded linear change of frame. Error bounds tie global hitting-time accuracy to one-step transition error. Their Isomorphic Embedding Learning algorithm anchors training with direct hitting-time regression, and maze locomotion tests improve on prior base policy results.

Catalog record

CREATORMagnus Victor Boock Abdullah Akgül Mustafa Mert Çelikok Melih Kandemir
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE133 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.06470

Browse the full library · About the Tower of Babel