Computer science · Paper · June 2022
Aakash Kumar
Convolutional Neural Networks (CNNs) are vulnerable to misclassifying images when small perturbations are present. With the increasing prevalence of CNNs in self-driving cars, it is vital to ensure these algorithms are robust to prevent collisions from occurring due to failure in recognizing a situation. In the Adversarial Self-Driving framework, a Generative Adversarial Network (GAN) is implemented to generate realistic perturbations in an image that cause a classifier CNN to misclassify data. This perturbed data is then used to train the classifier CNN further. The Adversarial Self-driving framework is applied to an image classification algorithm to improve the classification accuracy on perturbed images and is later applied to train a self-driving car to drive in a simulation. A small-scale self-driving car is also built to drive around a track and classify signs. The Adversarial Self-driving framework produces perturbed images through learning a dataset, as a result removing the need to train on significant amounts of data. Experiments demonstrate that the Adversarial Self-driving framework identifies situations where CNNs are vulnerable to perturbations and generates new examples of these situations for the CNN to train on. The additional data generated by the Adversarial Self-driving framework provides sufficient data for the CNN to generalize to the environment. Therefore, it is a viable tool to increase the resilience of CNNs to perturbations. Particularly, in the real-world self-driving car, the application of the Adversarial Self-Driving framework resulted in an 18% increase in accuracy, and the simulated self-driving model had no collisions in 30 minutes of driving.
Computer science · Paper · June 2022
Christopher Sun, Jai Sharma, Milind Maiti
Roughly 6,800 natural disasters occur annually, and this alarming number continues to grow due to climate change. Effective methods to improve natural disaster response include change detection, map alignment, and vision-aided navigation to allow for the time-efficient delivery of life-saving aid. Current software functions optimally only on near-nadir images taken around ninety degrees above ground level. The inability to generalize to oblique images increases the need to compute an image's geocentric pose, which is its spatial orientation with respect to gravity. This Deep Learning investigation presents three convolutional models to predict geocentric pose using 5,923 nadir and off-nadir red, green, and blue (RGB) satellite images of cities worldwide. Prior to the ensemble, an autoencoder is assembled to condense the 256 x 256 x 3 images to 32 x 32 x 16 latent space representations, demonstrating the ability to learn useful features from the data. The first model in the ensemble is a U-Net Fully Convolutional Network (FCN) with skip connections used to predict each image's corresponding pixel-level above-ground elevation mask. This FCN achieves a median absolute deviation of 0.335 meters and an R2 of 0.865 on test data. Afterward, the elevation masks are concatenated with the RGB images to form four-channel inputs fed into the second model in the ensemble, which predicts each image's rotation angle and magnification scale, the components of its geocentric pose. This Deep Convolutional Model achieves an R2 of 0.943 on test data, significantly outperforming previous models designed by researchers. In addition to achieving superior model performance, outlier removal was performed through supervised interpolation, and the usefulness of data features was gauged through a sensitivity analysis of elevation masks, to target future avenues of feature engineering. The high-accuracy software built in this study contributes to mapping and navigation procedures to accelerate disaster relief and save human lives.
Computer science · Paper · August 2021
Jai Sharma, Milind Maiti, Christopher Sun
Dropout Regularization, serving to reduce variance, is nearly ubiquitous in Deep Learning models. We explore the relationship between the dropout rate and model complexity by training 2,000 neural networks configured with random combinations of the dropout rate and the number of hidden units in each dense layer, on each of the three data sets we selected. The generated figures, with binary cross entropy loss and binary accuracy on the z-axis, question the common assumption that adding depth to a dense layer while increasing the dropout rate will certainly enhance performance. We also discover a complex correlation between the two hyperparameters that we proceed to quantify by building additional machine learning and Deep Learning models which predict the optimal dropout rate given some hidden units in each dense layer. Linear regression and polynomial logistic regression require the use of arbitrary thresholds to select the cost data points included in the regression and to assign the cost data points a binary classification, respectively. These machine learning models have mediocre performance because their naive nature prevented the modeling of complex decision boundaries. Turning to Deep Learning models, we build neural networks that predict the optimal dropout rate given the number of hidden units in each dense layer, the desired cost, and the desired accuracy of the model. Though, this attempt encounters a mathematical error that can be attributed to the failure of the vertical line test. The ultimate Deep Learning model is a neural network whose decision boundary represents the 2,000 previously generated data points. This final model leads us to devise a promising method for tuning hyperparameters to minimize computational expense yet maximize performance. The strategy can be applied to any model hyperparameters, with the prospect of more efficient tuning in industrial models.