SCNO composes pre-trained spiking neural operator blocks for elementary PDE terms to solve unseen coupled PDEs with a frozen library plus a lightweight correction network, achieving lower error than monolithic baselines using only 95K parameters.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
SCNO: Spiking Compositional Neural Operator -- Towards a Neuromorphic Foundation Model for Nuclear PDE Solving
SCNO composes pre-trained spiking neural operator blocks for elementary PDE terms to solve unseen coupled PDEs with a frozen library plus a lightweight correction network, achieving lower error than monolithic baselines using only 95K parameters.