Leonardo B. Oliveira

Artur Souza, Luigi Nardi, Leonardo B. Oliveira et al.

While Bayesian Optimization (BO) is a very popular method for optimizing expensive black-box functions, it fails to leverage the experience of domain experts. This causes BO to waste function evaluations on bad design choices (e.g., machine learning hyperparameters) that the expert already knows to work poorly. To address this issue, we introduce Bayesian Optimization with a Prior for the Optimum (BOPrO). BOPrO allows users to inject their knowledge into the optimization process in the form of priors about which parts of the input space will yield the best performance, rather than BO's standard priors over functions, which are much less intuitive for users. BOPrO then combines these priors with BO's standard probabilistic model to form a pseudo-posterior used to select which points to evaluate next. We show that BOPrO is around 6.67x faster than state-of-the-art methods on a common suite of benchmarks, and achieves a new state-of-the-art performance on a real-world hardware design application. We also show that BOPrO converges faster even if the priors for the optimum are not entirely accurate and that it robustly recovers from misleading priors.

Maria L. B. A. Santos, Jessica C. Carneiro, Antonio M. R. Franco et al.

Considering the world's IoT development and market, it is necessary to guarantee the security of the developed IoT applications as well as the privacy of their end users. In this sense, Federated Identity Management (FIdM) systems can be of great help as they improve user authentication and privacy. In this paper, we claim that traditional FIdM are mostly cumbersome and then ill-suited for IoT. As a solution to this problem, we come up with a federated identity authentication protocol exclusively tailored to IoT. Federated Lightweight Authentication of Things (FLAT), our solution, replaces weighty protocols and asymmetric cryptographic primitives used in traditional FIdM by lighter ones. For instance, FLAT synergistically combines symmetric cryptosystems and Implicit Certificates. The results show that FLAT can reduce the data exchange overhead by around 31% when compared to a baseline solution. FLAT's Client is also more efficient than the baseline solution in terms of data transmitted, data received, total data exchange, and computation time. Our results indicate that FLAT runs efficiently even on top of resource-constrained devices like Arduino.

Artur Souza, Antônio A. F. Loureiro, Leonardo B. Oliveira

We quickly approach a "pervasive future" where pervasive computing is the norm. In this scenario, humans are surrounded by a multitude of heterogeneous devices that assist them in almost every aspect of their daily routines. The realization of this future demands strong authentication guarantees to ensure that these devices are not abused and that their users are not endangered. However, providing authentication for these systems is a challenging task due to the high heterogeneity of pervasive computing applications. This heterogeneity makes it unfeasible to propose a single authentication solution for all of the pervasive computing applications. In this paper, we review several pervasive application scenarios and promising authentication methods for each. To do this, we first identify the key characteristics of each pervasive application scenario. Then, we review the strengths and weaknesses of prominent authentication methods from the literature. Finally, we identify which authentication methods are well suited for each application scenario based on the identified characteristics. Our goal is to provide promising directions to be explored for authentication in each of these scenarios.

Artur Souza, Leonardo B. Oliveira, Sabine Hollatz et al.

Serial crystallography is the field of science that studies the structure and properties of crystals via diffraction patterns. In this paper, we introduce a new serial crystallography dataset comprised of real and synthetic images; the synthetic images are generated through the use of a simulator that is both scalable and accurate. The resulting dataset is called DiffraNet, and it is composed of 25,457 512x512 grayscale labeled images. We explore several computer vision approaches for classification on DiffraNet such as standard feature extraction algorithms associated with Random Forests and Support Vector Machines but also an end-to-end CNN topology dubbed DeepFreak tailored to work on this new dataset. All implementations are publicly available and have been fine-tuned using off-the-shelf AutoML optimization tools for a fair comparison. Our best model achieves 98.5% accuracy on synthetic images and 94.51% accuracy on real images. We believe that the DiffraNet dataset and its classification methods will have in the long term a positive impact in accelerating discoveries in many disciplines, including chemistry, geology, biology, materials science, metallurgy, and physics.