]> 2026-08-26T14:32:50+00:00 Inverse Abstraction of Neural Networks Using Symbolic Interpolation 0 en Neural networks in real-world applications have to satisfy critical properties such as safety and reliability. The analysis of such properties typically involves extracting informa- tion through computing pre-images of neural networks, but it is well-known that explicit computation of pre-images is intractable. We introduce new methods for computing compact symbolic abstractions of pre-images. Our approach relies on computing approximations that provably overapproximate and underapproximate the pre-images at all layers. The abstraction of pre-images enables formal analysis and knowl- edge extraction without modifying standard learning algo- rithms. We show how to use inverse abstractions to automatically extract simple control laws and compact representations for pre-images corresponding to unsafe outputs. We illustrate that the extracted abstractions are often interpretable and can be used for analyzing complex properties. Sumanth Dathathri Sicun Gao Richard M. Murray 2018e To appear, 2019 AAAI Conference on Artificial Intelligence dgm19-aiaa Conference paper 2018-12-27T05:21:34Z 2458479.7233102 Inverse Abstraction of Neural Networks Using Symbolic Interpolation Title 102 en Title