Search

Although rare diseases are characterized by low prevalence, approximately 300 million people are affected by a rare disease. The early and accurate diagnosis of these conditions is a major challenge for general practitioners, who do not have enough knowledge to identify them. In addition to this, rare diseases usually show a wide variety of manifestations, which might make the diagnosis even more difficult. A delayed diagnosis can negatively affect the patient's life. Therefore, there is an urgent need to increase the scientific and medical knowledge about rare diseases. Natural Language Processing (NLP) and Deep Learning can help to extract relevant information about rare diseases to facilitate their diagnosis and treatments. The paper explores the use of several deep learning techniques such as Bidirectional Long Short Term Memory (BiLSTM) networks or deep contextualized word representations based on Bidirectional Encoder Representations from Transformers (BERT) to recognize rare diseases and their clinical manifestations (signs and symptoms) in the RareDis corpus. This corpus contains more than 5,000 rare diseases and almost 6,000 clinical manifestations. BioBERT, a domain-specific language representation based on BERT and trained on biomedical corpora, obtains the best results. In particular, this model obtains an F1-score of 85.2% for rare diseases, outperforming all the other models.

Based on the Global Entrepreneurship Monitor (GEM) surveys and conducting a panel data estimation to test our hypothesis, this paper examines whether corruption perceptions might sand or grease the wheels for entrepreneurship inside companies or intrapreneurship in a sample of 92 countries for the period 2012 to 2019. Our results find that the corruption perception sands the wheel for intrapreneurship. There is evidence of a quadratic relation, but this relation is only clear for the less developed countries, which sort of moderate the very negative effect of corruption for these countries. The results also confirm that corruption influences differently on intrapreneurship depending on the level of development of the country., Comment: 18 pages

The motion control of wheeled mobile robots at high speeds under adverse ground conditions is a difficult task, since the robots' wheels may be subject to different kinds of slip. This work introduces an adaptive kinematic controller that is capable of solving the trajectory tracking problem of a nonholonomic mobile robot under longitudinal and lateral slip. While the controller can effectively compensate for the longitudinal slip, the lateral slip is a more involved problem to deal with, since nonholonomic robots cannot directly produce movement in the lateral direction. To show that the proposed controller is still able to make the mobile robot follow a reference trajectory under lateral and longitudinal time-varying slip, the solutions of the robot's position and orientation error dynamics are shown to be uniformly ultimately bounded. Numerical simulations are presented to illustrate the robot's performance using the proposed adaptive control law.

Current Continuous Integration processes face significant intrinsic cybersecurity challenges. The idea is not only to solve and test formal or regulatory security requirements of source code but also to adhere to the same principles to the CI pipeline itself. This paper presents an overview of current security issues in CI workflow. It designs, develops, and deploys a new tool for the secure deployment of a container-based CI pipeline flow without slowing down release cycles. The tool, called \SD for its Docker-based approach, is publicly available in GitHub. It implements a transparent application firewall based on a configuration mechanism avoiding issues in the CI workflow associated with intended or unintended container configurations. Integrated with other DevOps Engineers tools, it provides feedback from only those scenarios that match specific patterns, addressing future container security issues., Comment: Preprint: 15 pages, 5 figures, 2 tables. Public repository: https://github.com/uleroboticsgroup/Secdocker

Ethereum prospered the inception of a plethora of smart contract applications, ranging from gambling games to decentralized finance. However, Ethereum is also considered a highly adversarial environment, where vulnerable smart contracts will eventually be exploited. Recently, Ethereum's pool of pending transaction has become a far more aggressive environment. In the hope of making some profit, attackers continuously monitor the transaction pool and try to frontrun their victims' transactions by either displacing or suppressing them, or strategically inserting their transactions. This paper aims to shed some light into what is known as a dark forest and uncover these predators' actions. We present a methodology to efficiently measure the three types of frontrunning: displacement, insertion, and suppression. We perform a large-scale analysis on more than 11M blocks and identify almost 200K attacks with an accumulated profit of 18.41M USD for the attackers, providing evidence that frontrunning is both, lucrative and a prevalent issue.

The idea of employing non-Abelian statistics for error-free quantum computing ignited interest in recent reports of topological surface superconductivity and Majorana zero modes (MZMs) in FeTe$_{0.55}$Se$_{0.45}$. An associated puzzle is that the topological features and superconducting properties are not observed uniformly across the sample surface. Understanding and practical control of these electronic inhomogeneities present a prominent challenge for potential applications. Here, we combine neutron scattering, scanning angle-resolved photoemission spectroscopy (ARPES), and microprobe composition and resistivity measurements to characterize the electronic state of Fe$_{1+y}$Te$_{1-x}$Se$_{x}$. We establish a phase diagram in which the superconductivity is observed only at sufficiently low Fe concentration, in association with distinct antiferromagnetic correlations, while the coexisting topological surface state occurs only at sufficiently high Te concentration. We find that FeTe$_{0.55}$Se$_{0.45}$ is located very close to both phase boundaries, which explains the inhomogeneity of superconducting and topological states. Our results demonstrate the compositional control required for use of topological MZMs in practical applications., Comment: Main text (18 pages, 5 figures, 50 references) combined with Supplementary (18 pages, 12 figures). Nat. Mater. (2021)

Social choice is the theory about collective decision towards social welfare starting from individual opinions, preferences, interests or welfare. The field of Computational Social Welfare is somewhat recent and it is gaining impact in the Artificial Intelligence Community. Classical literature makes the assumption of single-peaked preferences, i.e. there exist a order in the preferences and there is a global maximum in this order. This year some theoretical results were published about Two-stage Approval Voting Systems (TAVs), Multi-winner Selection Rules (MWSR) and Incomplete (IPs) and Circular Preferences (CPs). The purpose of this paper is three-fold: Firstly, I want to introduced Social Choice Optimisation as a generalisation of TAVs where there is a max stage and a min stage implementing thus a Minimax, well-known Artificial Intelligence decision-making rule to minimize hindering towards a (Social) Goal. Secondly, I want to introduce, following my Open Standardization and Open Integration Theory (in refinement process) put in practice in my dissertation, the Open Standardization of Social Inclusion, as a global social goal of Social Choice Optimization.

In practice, machine learning experts are often confronted with imbalanced data. Without accounting for the imbalance, common classifiers perform poorly and standard evaluation metrics mislead the practitioners on the model's performance. A common method to treat imbalanced datasets is under- and oversampling. In this process, samples are either removed from the majority class or synthetic samples are added to the minority class. In this paper, we follow up on recent developments in deep learning. We take proposals of deep generative models, including our own, and study the ability of these approaches to provide realistic samples that improve performance on imbalanced classification tasks via oversampling. Across 160K+ experiments, we show that all of the new methods tend to perform better than simple baseline methods such as SMOTE, but require different under- and oversampling ratios to do so. Our experiments show that the way the method of sampling does not affect quality, but runtime varies widely. We also observe that the improvements in terms of performance metric, while shown to be significant when ranking the methods, often are minor in absolute terms, especially compared to the required effort. Furthermore, we notice that a large part of the improvement is due to undersampling, not oversampling. We make our code and testing framework available.

We study the matching problem of regular tree languages, that is, "$\exists \sigma:\sigma(L)\subseteq R$?" where $L,R$ are regular tree languages over the union of finite ranked alphabets $\Sigma$ and $\mathcal{X}$ where $\mathcal{X}$ is an alphabet of variables and $\sigma$ is a substitution such that $\sigma(x)$ is a set of trees in $T(\Sigma\cup H)\setminus H$ for all $x\in \mathcal{X}$. Here, $H$ denotes a set of "holes" which are used to define a "sorted" concatenation of trees. Conway studied this problem in the special case for languages of finite words in his classical textbook "Regular algebra and finite machines" published in 1971. He showed that if $L$ and $R$ are regular, then the problem "$\exists \sigma \forall x\in \mathcal{X}: \sigma(x)\neq \emptyset\wedge \sigma(L)\subseteq R$?" is decidable. Moreover, there are only finitely many maximal solutions, the maximal solutions are regular substitutions, and they are effectively computable. We extend Conway's results when $L,R$ are regular languages of finite and infinite trees, and language substitution is applied inside-out, in the sense of Engelfriet and Schmidt (1977/78). More precisely, we show that if $L\subseteq T(\Sigma\cup\mathcal{X})$ and $R\subseteq T(\Sigma)$ are regular tree languages over finite or infinite trees, then the problem "$\exists \sigma \forall x\in \mathcal{X}: \sigma(x)\neq \emptyset\wedge \sigma_{\mathrm{io}}(L)\subseteq R$?" is decidable. Here, the subscript "$\mathrm{io}$" in $\sigma_{\mathrm{io}}(L)$ refers to "inside-out". Moreover, there are only finitely many maximal solutions $\sigma$, the maximal solutions are regular substitutions and effectively computable. The corresponding question for the outside-in extension $\sigma_{\mathrm{oi}}$ remains open, even in the restricted setting of finite trees.

Despite mounting evidence that materials imperfections are a major obstacle to practical applications of superconducting qubits, connections between microscopic material properties and qubit coherence are poorly understood. Here, we perform measurements of transmon qubit relaxation times $T_1$ in parallel with spectroscopy and microscopy of the thin polycrystalline niobium films used in qubit fabrication. By comparing results for films deposited using three techniques, we reveal correlations between $T_1$ and grain size, enhanced oxygen diffusion along grain boundaries, and the concentration of suboxides near the surface. Physical mechanisms connect these microscopic properties to residual surface resistance and $T_1$ through losses arising from the grain boundaries and from defects in the suboxides. Further, experiments show that the residual resistance ratio can be used as a figure of merit for qubit lifetime. This comprehensive approach to understanding qubit decoherence charts a pathway for materials-driven improvements of superconducting qubit performance.

The implementation of aberration-corrected electron beam lithography (AC-EBL) in a 200 keV scanning transmission electron microscope (STEM) is a novel technique that could be used for the fabrication of quantum devices based on 2D atomic crystals with single nanometer critical dimensions, allowing to observe more robust quantum effects. In this work we study electron beam sculpturing of nanostructures on suspended graphene field effect transistors using AC-EBL, focusing on the in situ characterization of the impact of electron beam exposure on device electronic transport quality. When AC-EBL is performed on a graphene channel (local exposure) or on the outside vicinity of a graphene channel (non-local exposure), the charge transport characteristics of graphene can be significantly affected due to charge doping and scattering. While the detrimental effect of non-local exposure can be largely removed by vigorous annealing, local-exposure induced damage is irreversible and cannot be fixed by annealing. We discuss the possible causes of the observed exposure effects. Our results provide guidance to the future development of high-energy electron beam lithography for nanomaterial device fabrication.