AXIOM: Learning to Play Games in Minutes with Expanding Object-Centric Models
Conor Heins, Toon Van de Maele, Alexander Tschantz, Hampus Linander, Dimitrije Markovic, Tommaso Salvatori, Corrado Pezzato, Ozan Catal, Ran Wei, Magnus Koudahl, Marco Perin, Karl Friston, Tim Verbelen, Christopher L Buckley · 2025
Read AXIOM: Learning to Play Games in Minutes with Expanding Object-Centric Models by Conor Heins, Toon Van de Maele, Alexander Tschantz, Hampus Linander, Dimitrije Markovic, Tommaso Salvatori, Corrado Pezzato, Ozan Catal, Ran Wei, Magnus Koudahl, Marco Perin, Karl Friston, Tim Verbelen, Christopher L Buckley (2025) 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-2505-24784.txt.
Description
Conor Heins and co-authors built in 2025 a game learner, AXIOM, that masters games in minutes by modeling the world as objects. Deep reinforcement learners are data hungry, while people learn fast from built-in hunches about objects. Active inference gives a principled way to use such hunches, but past builds were hand-made per task. AXIOM carries a small, general kit of object priors across games.
Scenes split into objects whose paths run as straight-line pieces with sparse interplay. The model grows itself online, sprouting mixture parts from single events and pruning by Bayesian rules. Within ten thousand steps it masters varied games with few parameters and no gradient training. The team pitches it as a bridge between Bayesian thrift and deep learning's range.
Catalog record
| CREATOR | Conor Heins, Toon Van de Maele, Alexander Tschantz, Hampus Linander, Dimitrije Markovic, Tommaso Salvatori, Corrado Pezzato, Ozan Catal, Ran Wei, Magnus Koudahl, Marco Perin, Karl Friston, Tim Verbelen, Christopher L Buckley |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 136 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-24784