Martin Lange
2 indexed papers
Publications per year
Top categories
Frequent co-authors
Research Timeline
This paper analyzes the computational complexity of verifying feedforward neural networks when their weights are restricted to finite-width arithmetic, finding that verification remains NP-complete for fixed-precision networks and establishing upper bounds for dynamically quantized networks.
This paper analyzes the computational complexity of verifying feedforward neural networks when their weights are restricted to finite-width arithmetic, finding that verification remains NP-complete for fixed-precision networks and establishing upper bounds for dynamically quantized networks.
Papers
The Complexity of Verifying Feedforward Neural Networks in Quantised Settings
This paper analyzes the computational complexity of verifying feedforward neural networks when their weights are restricted to finite-width arithmetic, finding that verification remains NP-complete fo…