The authors present a Generative Adversarial Network (GAN) model that learns how to generate 3D models in their native format so that they can either be evaluated using complex simulation environments, or realized using methods such as additive manufacturing. Once initially trained, the GAN can create additional training data itself by generating new designs, evaluating them in a physics-based virtual environment, and adding the high performing ones to the training set. A case study involving a GAN model that is initially trained on 4045 3D aircraft models is used for demonstration, where the training data set updated with GAN-generated and evaluated designs, results in enhanced model generation, both in the geometric feasibility and performance of the designs. Z-tests on the performance scores of the generated aircraft models indicate a statistically significant improvement in the functionality of the generated models after three iterations of the training-evaluation process. In the case study, a number of techniques are explored to structure the generate-evaluate process in order to balance the need to generate feasible designs with the need for innovative designs.

No abstract

Intrusion Detection Systems (IDS) are increasingly adopting machine learning (ML)-based approaches to detect threats in computer networks due to their ability to learn underlying threat patterns/features. However, ML-based models are susceptible to adversarial attacks, attacks wherein slight perturbations of the input features, cause misclassifications. We propose a method that uses active learning and generative adversarial networks to evaluate the threat of adversarial attacks on ML-based IDS. Existing adversarial attack methods require a large amount of training data or assume knowledge of the IDS model itself (e.g., loss function), which may not be possible in real-world settings. Our method overcomes these limitations by demonstrating the ability to compromise an IDS using limited training data and assuming no prior knowledge of the IDS model other than its binary classification (i.e., benign or malicious). Experimental results demonstrate the ability of our proposed model to achieve a 98.86% success rate in bypassing the IDS model using only 25 labeled data points during model training. The knowledge gained by compromising the ML-based IDS, can be integrated into the IDS in order to enhance its robustness against similar ML-based adversarial attacks. CCS CONCEPTS• Security and privacy → Intrusion/anomaly detection and malware mitigation;

No abstract

Generative neural networks (GNNs) have successfully used human-created designs to generate novel 3D models that combine concepts from disparate known solutions, which is an important aspect of design exploration. GNNs automatically learn a parameterization (or latent space) of a design space, as opposed to alternative methods that manually define a parameterization. However, GNNs are typically not evaluated using an explicit notion of physical performance, which is a critical capability needed for design. This work bridges this gap by proposing a method to extract a set of functional designs from the latent space of a point cloud generating GNN, without sacrificing the aforementioned aspects of a GNN that are appealing for design exploration. We introduce a sparsity preserving cost function and initialization strategy for a genetic algorithm (GA) to optimize over the latent space of a point cloud generating autoencoder GNN. We examine two test cases, an example of generating ellipsoid point clouds subject to a simple performance criterion and a more complex example of extracting 3D designs with a low coefficient of drag. Our experiments show that the modified GA results in a diverse set of functionally superior designs while maintaining similarity to human-generated designs in the training data set.