Prophecy infers formal properties of feed-forward neural networks by extracting rules from neuron activation patterns that imply desirable output behaviors.
Runtime enforcement using knowledge bases
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2025 2verdicts
UNVERDICTED 2representative citing papers
The authors define semantic lifting and semantic reflection for the small language SMOL, formalize its operational semantics and type system, and demonstrate it on a geological modelling case study with an open-source implementation.
citing papers explorer
-
Prophecy: Inferring Formal Properties from Neuron Activations
Prophecy infers formal properties of feed-forward neural networks by extracting rules from neuron activation patterns that imply desirable output behaviors.
-
Semantically Reflected Programs
The authors define semantic lifting and semantic reflection for the small language SMOL, formalize its operational semantics and type system, and demonstrate it on a geological modelling case study with an open-source implementation.