SPAN, a KHRONOS-derived spline network with a learnable preprocessing layer, reports better sample efficiency and success rates than small MLPs across several RL benchmarks, though the abstract overclaims and omits the HVAC results.
URL https://arc.aiaa
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Agile Reinforcement Learning through Separable Neural Architecture and Applications
SPAN, a KHRONOS-derived spline network with a learnable preprocessing layer, reports better sample efficiency and success rates than small MLPs across several RL benchmarks, though the abstract overclaims and omits the HVAC results.