Generative Neural Networks (GNN) have demonstrated remarkable power in creating novel graphic design images from text-to-image training. This work applies GNNs to design dynamic inductive power transfer systems to make charging electrical vehicles (EVs) more convenient and cheaper. Discovering optimal and safe coil implementations (for EV and road) is challenging because of the combinatorial explosion of possible configurations, along with multiple conflicting objective functions, such as maximizing the output power, while minimizing stray magnetic fields and the volume of windings and magnetic cores. To solve the problem, a differentiable simulator and evaluator define loss functions that train a generative neural network to only produce configurations that satisfy eight given design criteria. Before training, the rate of finding successful designs is 0.005%, but within 500 training epochs, the rate becomes 98% (about 30 seconds run time). Solution diversity and quality may be further improved by applying loss functions trained on pairwise Pareto-optimal generated examples.

The recent eminent success of deep learning models has triggered a lot of interest in data augmentation studies. The main challenge of data-hungry models is that they are bound to overfitting and require massive amounts of data to be trained on to achieve optimal levels of convergence. It is of high importance for any data oversampling algorithm to produce synthetic data samples that are not only stemming from the same real data distribution, but also have enough variation to provide more learning opportunities to trained models. Early efforts on time series data augmentation either rely on generating new samples by interpolating between two close real data neighbors or using Generative Adversarial Networks. In this paper, we present Time Series Generation based on Auto-encoders (TiGA), a novel algorithm for time series generation using time-warped autoencoders. Our idea is to exploit the lossy transformation of autoencoders for the purpose of generating synthetic time series data samples. To the best of our knowledge, this is the first effort that leverages the latent features generated by autoencoders for the purpose of time series data oversampling. We evaluate our proposed approach on an open-source real-life solar flare prediction dataset. Results show that TiGA produces samples that are both quantitatively and qualitatively superior to current state-of-the-art methods.