Abstract:Formal verification is only as good as the specification of a system, which is also true for neural network verification. Existing specifications follow the paradigm of data as specification, where the local neighborhood around a reference data point is considered correct or robust. While these specifications provide a fair testbed for assessing model robustness, they are too restrictive for verifying unseen test data-a challenging task with significant real-world implications. Recent work shows great promise through a new paradigm, neural representation as specification, which uses neural activation patterns (NAPs) for this purpose. However, it computes the most refined NAPs, which include many redundant neurons. In this paper, we study the following problem: Given a neural network, find a minimal (general) NAP specification that is sufficient for formal verification of the network's robustness. Finding the minimal NAP specification not only expands verifiable bounds but also provides insights into which neurons contribute to the model's robustness. To address this problem, we propose several exact and approximate approaches. Our exact approaches leverage the verification tool to find minimal NAP specifications in either a deterministic or statistical manner. Whereas the approximate methods efficiently estimate minimal NAPs using adversarial examples and local gradients, without making calls to the verification tool. This allows us to inspect potential causal links between neurons and the robustness of state-of-the art neural networks, a task for which existing verification frameworks fail to scale. Our experimental results suggest that minimal NAP specifications require much smaller fractions of neurons compared to the most refined NAP specifications computed by previous work, yet they can significantly expand the verifiable boundaries to several orders of magnitude larger.

A Minimal Neural Network Ensemble Construction Method: A Constructive Approach

A Neural Network Model For General Minimax Problem

A Constructive Algorithm for Training Neural Network Ensemble

Be expert in multiple aspects and good at many modular neural network with reduced subnet relative training complexity

Neural Subnetwork Ensembles

Ensemble Neural Networks (ENN): A gradient-free stochastic method

SAE: Single Architecture Ensemble Neural Networks

Learning Minimal Neural Specifications

A DNN-Ensemble Method for Error Reduction and Training Data Selection in DNN Based Modeling

Searching for Minimal Optimal Neural Networks

Sequential Bayesian Neural Subnetwork Ensembles

The Diversified Ensemble Neural Network.

Research and development of neural network ensembles: a survey

Maximum-and-Concatenation Networks

A fragmented neural network ensemble method and its application to image classification

k-Winners-Take-All Ensemble Neural Network

Max and Coincidence Neurons in Neural Networks

Nucleus Neural Network: A Data-driven Self-organized Architecture

Genetic Algorithm based Selective Neural Network Ensemble.

One-Shot Neural Ensemble Architecture Search by Diversity-Guided Search Space Shrinking