An agent that selects proto-objects (segmented color regions) instead of fixed patches matches or beats a patch-based baseline on CarRacing and Doom Take Cover while using 62% fewer parameters in its smallest configuration.
Finkel and Paul Sajda
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.NE 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Neuroevolution of Self-Attention Over Proto-Objects
An agent that selects proto-objects (segmented color regions) instead of fixed patches matches or beats a patch-based baseline on CarRacing and Doom Take Cover while using 62% fewer parameters in its smallest configuration.