Towards robust autonomous driving systems through adversarial test set generation.

Correct environmental perception of objects on the road is vital for the safety of autonomous driving. Making appropriate decisions by the autonomous driving algorithm could be hindered by data perturbations and more recently, by adversarial attacks. We propose an adversarial test input generation approach based on uncertainty to make the machine learning (ML) model more robust against data perturbations and adversarial attacks. Adversarial attacks and uncertain inputs can affect the ML model's performance, which can have severe consequences such as the misclassification of objects on the road by autonomous vehicles, leading to incorrect decision-making. We show that we can obtain more robust ML models for autonomous driving by making a dataset that includes highly-uncertain adversarial test inputs during the re-training phase. We demonstrate an improvement in the accuracy of the robust model by more than 12%, with a notable drop in the uncertainty of the decisions returned by the model. We believe our approach will assist in further developing risk-aware autonomous systems.

Systematic Identification of Atom-Centered Symmetry Functions for the Development of Neural Network Potentials.

Neural network potentials are emerging as promising classical force fields that can enable long-time and large-length scale simulations at close to ab initio accuracies. They learn the underlying potential energy surface by mapping the Cartesian coordinates of atoms to system energies using elemental neural networks. To ensure invariance with respect to system translation, rotation, and atom index permutations, in the Behler-Parrinnello type of neural network potential (BP-NNP), the Cartesian coordinates of atoms are transformed into "structural fingerprints" using atom-centered symmetry functions (ACSFs). Development of an accurate BP-NNP for any chemical system critically relies on the choice of these ACSFs. In this work, we have proposed a systematic framework for the identification of an optimal set of ACSFs for any target system, which not only considers the diverse atomic environments present in the training dataset but also inter-ACSF correlations. Our method is applicable to different kinds of ACSFs and across diverse chemical systems. We demonstrate this by building accurate BP-NNPs for water and Cu2S systems.

Time-series forecasting of Bitcoin prices using high-dimensional features: a machine learning approach.

Bitcoin is a decentralized cryptocurrency, which is a type of digital asset that provides the basis for peer-to-peer financial transactions based on blockchain technology. One of the main problems with decentralized cryptocurrencies is price volatility, which indicates the need for studying the underlying price model. Moreover, Bitcoin prices exhibit non-stationary behavior, where the statistical distribution of data changes over time. This paper demonstrates high-performance machine learning-based classification and regression models for predicting Bitcoin price movements and prices in short and medium terms. In previous works, machine learning-based classification has been studied for an only one-day time frame, while this work goes beyond that by using machine learning-based models for one, seven, thirty and ninety days. The developed models are feasible and have high performance, with the classification models scoring up to 65% accuracy for next-day forecast and scoring from 62 to 64% accuracy for seventh-ninetieth-day forecast. For daily price forecast, the error percentage is as low as 1.44%, while it varies from 2.88 to 4.10% for horizons of seven to ninety days. These results indicate that the presented models outperform the existing models in the literature.

Suitability of the Openly Accessible 3D Printed Prosthetic Hands for War-Wounded Children.

The field of rehabilitation and assistive devices is being disrupted by innovations in desktop 3D printers and open-source designs. For upper limb prosthetics, those technologies have demonstrated a strong potential to aid those with missing hands. However, there are basic interfacing issues that need to be addressed for long term usage. The functionality, durability, and the price need to be considered especially for those in difficult living conditions. We evaluated the most popular designs of body-powered, 3D printed prosthetic hands. We selected a representative sample and evaluated its suitability for its grasping postures, durability, and cost. The prosthetic hand can perform three grasping postures out of the 33 grasps that a human hand can do. This corresponds to grasping objects similar to a coin, a golf ball, and a credit card. Results showed that the material used in the hand and the cables can withstand a 22 N normal grasping force, which is acceptable based on standards for accessibility design. The cost model showed that a 3D printed hand could be produced for as low as $19. For the benefit of children with congenital missing limbs and for the war-wounded, the results can serve as a baseline study to advance the development of prosthetic hands that are functional yet low-cost.