Search

While deep face recognition models have demonstrated remarkable performance, they often struggle on the inputs from domains beyond their training data. Recent attempts aim to expand the training set by relying on computationally expensive and inherently challenging image-space augmentation of image generation modules. In an orthogonal direction, we present a simple yet effective method to expand the training data by interpolating between instance-level feature statistics across labeled and unlabeled sets. Our method, dubbed Targeted Style Adversary (TSA), is motivated by two observations: (i) the input domain is reflected in feature statistics, and (ii) face recognition model performance is influenced by style information. Shifting towards an unlabeled style implicitly synthesizes challenging training instances. We devise a recognizability metric to constraint our framework to preserve the inherent identity-related information of labeled instances. The efficacy of our method is demonstrated through evaluations on unconstrained benchmarks, outperforming or being on par with its competitors while offering nearly a 70\% improvement in training speed and 40\% less memory consumption.

We build interpretable and lightweight transformer-like neural networks by unrolling iterative optimization algorithms that minimize graph smoothness priors -- the quadratic graph Laplacian regularizer (GLR) and the $\ell_1$-norm graph total variation (GTV) -- subject to an interpolation constraint. The crucial insight is that a normalized signal-dependent graph learning module amounts to a variant of the basic self-attention mechanism in conventional transformers. Unlike "black-box" transformers that require learning of large key, query and value matrices to compute scaled dot products as affinities and subsequent output embeddings, resulting in huge parameter sets, our unrolled networks employ shallow CNNs to learn low-dimensional features per node to establish pairwise Mahalanobis distances and construct sparse similarity graphs. At each layer, given a learned graph, the target interpolated signal is simply a low-pass filtered output derived from the minimization of an assumed graph smoothness prior, leading to a dramatic reduction in parameter count. Experiments for two image interpolation applications verify the restoration performance, parameter efficiency and robustness to covariate shift of our graph-based unrolled networks compared to conventional transformers.

Collision-free motion is essential for mobile robots. Most approaches to collision-free and efficient navigation with wheeled robots require parameter tuning by experts to obtain good navigation behavior. This study investigates the application of deep reinforcement learning to train a mobile robot for autonomous navigation in a complex environment. The robot utilizes LiDAR sensor data and a deep neural network to generate control signals guiding it toward a specified target while avoiding obstacles. We employ two reinforcement learning algorithms in the Gazebo simulation environment: Deep Deterministic Policy Gradient and proximal policy optimization. The study introduces an enhanced neural network structure in the Proximal Policy Optimization algorithm to boost performance, accompanied by a well-designed reward function to improve algorithm efficacy. Experimental results conducted in both obstacle and obstacle-free environments underscore the effectiveness of the proposed approach. This research significantly contributes to the advancement of autonomous robotics in complex environments through the application of deep reinforcement learning., Comment: This paper is under review by Int. J. of Intelligent Machines and Robotics

We study joint optimization of service placement, request routing, and CPU sizing in a cooperative MEC system. The problem is considered from the perspective of the service provider (SP), which delivers heterogeneous MEC-enabled delay-sensitive services, and needs to pay for the used resources to the mobile network operators and the cloud provider, while earning revenue from the served requests. We formulate the problem of maximizing the SP's total profit subject to the computation, storage, and communication constraints of each edge node and end-to-end delay requirements of the services as a mixed-integer non-convex optimization problem, and prove it to be NP-hard. To tackle the challenges in solving the problem, we first introduce a design trade-off parameter for different delay requirements of each service, which maintains flexibility in prioritizing them, and transform the original optimization problem by the new delay constraints. Then, by exploiting a hidden convexity, we reformulate the delay constraints into an equivalent form. Next, to handle the challenge of the complicating (integer) variables, using primal decomposition, we decompose the problem into an equivalent form of master and inner sub-problems over the mixed and real variables, respectively. We then employ a cutting-plane approach for building up adequate representations of the extremal value of the inner problem as a function of the complicating variables and the set of values of the complicating variables for which the inner problem is feasible. Finally, we propose a solution strategy based on generalized Benders decomposition and prove its convergence to the optimal solution within a limited number of iterations. Extensive simulation results demonstrate that the proposed scheme significantly outperforms the existing mechanisms in terms of the SP's profit, cache hit ratio, running time, and end-to-end delay.

We present a new approach to machine learning-powered combinatorial auctions, which is based on the principles of Differential Privacy. Our methodology guarantees that the auction mechanism is truthful, meaning that rational bidders have the incentive to reveal their true valuation functions. We achieve this by inducing truthfulness in the auction dynamics, ensuring that bidders consistently provide accurate information about their valuation functions. Our method not only ensures truthfulness but also preserves the efficiency of the original auction. This means that if the initial auction outputs an allocation with high social welfare, our modified truthful version of the auction will also achieve high social welfare. We use techniques from Differential Privacy, such as the Exponential Mechanism, to achieve these results. Additionally, we examine the application of differential privacy in auctions across both asymptotic and non-asymptotic regimes.

This paper discusses the results of the third edition of the Monocular Depth Estimation Challenge (MDEC). The challenge focuses on zero-shot generalization to the challenging SYNS-Patches dataset, featuring complex scenes in natural and indoor settings. As with the previous edition, methods can use any form of supervision, i.e. supervised or self-supervised. The challenge received a total of 19 submissions outperforming the baseline on the test set: 10 among them submitted a report describing their approach, highlighting a diffused use of foundational models such as Depth Anything at the core of their method. The challenge winners drastically improved 3D F-Score performance, from 17.51% to 23.72%., Comment: To appear in CVPRW2024

We provide a first principles derivation of the microscopic entropy of a very general class of supersymmetric, rotating and accelerating black holes in AdS$_4$. This is achieved by analysing the large-$N$ limit of the spindle index and completes the construction of the first example of a holographic duality involving supersymmetric field theories defined on orbifolds with conical singularities., Comment: 6 pages, v2: references added

In 2019, the world faced a new challenge: a COVID-19 disease caused by the novel coronavirus, SARS-CoV-2. The virus rapidly spread across the globe, leading to a high rate of mortality, which prompted health organizations to take measures to control its transmission. Early disease detection is crucial in the treatment process, and computer-based automatic detection systems have been developed to aid in this effort. These systems often rely on artificial intelligence (AI) approaches such as machine learning, neural networks, fuzzy systems, and deep learning to classify diseases. This study aimed to differentiate COVID-19 patients from others using self-categorizing classifiers and employing various AI methods. This study used two datasets: the blood test samples and radiography images. The best results for the blood test samples obtained from San Raphael Hospital, which include two classes of individuals, those with COVID-19 and those with non-COVID diseases, were achieved through the use of the Ensemble method (a combination of a neural network and two machines learning methods). The results showed that this approach for COVID-19 diagnosis is cost-effective and provides results in a shorter amount of time than other methods. The proposed model achieved an accuracy of 94.09% on the dataset used. Secondly, the radiographic images were divided into four classes: normal, viral pneumonia, ground glass opacity, and COVID-19 infection. These were used for segmentation and classification. The lung lobes were extracted from the images and then categorized into specific classes. We achieved an accuracy of 91.1% on the image dataset. Generally, this study highlights the potential of AI in detecting and managing COVID-19 and underscores the importance of continued research and development in this field., Comment: This paper is under review by Int. J. of Computational Science and Engineering

Lane detection for autonomous vehicles is an important concept, yet it is a challenging issue of driver assistance systems in modern vehicles. The emergence of deep learning leads to significant progress in self-driving cars. Conventional deep learning-based methods handle lane detection problems as a binary segmentation task and determine whether a pixel belongs to a line. These methods rely on the assumption of a fixed number of lanes, which does not always work. This study aims to develop an optimal structure for the lane detection problem, offering a promising solution for driver assistance features in modern vehicles by utilizing a machine learning method consisting of binary segmentation and Affinity Fields that can manage varying numbers of lanes and lane change scenarios. In this approach, the Convolutional Neural Network (CNN), is selected as a feature extractor, and the final output is obtained through clustering of the semantic segmentation and Affinity Field outputs. Our method uses less complex CNN architecture than existing ones. Experiments on the TuSimple dataset support the effectiveness of the proposed method., Comment: The paper is under review by Int. J. of Mechatronics and Automation

New strategy is presented to optimize the performance of Thermoelectric (TE) coolers. This approach breaks optimizing TE coolers free from traditional methods of controlling temperature or engineering materials and the structural properties of the junctions. We introduced a dimensionless figure, {\gamma}, that shows the ratio of the unavailable cooling capacity to the available cooling capacity. This parameter relates the TE coolers' coefficient of performance (COP) to the COP of the reversible cycle (second law of thermodynamics efficiency) for a given electrical current. The theoretical description of the model is presented, and it is shown that controlling {\gamma} during the TE performance minimizes entropy generation and energy loss, which leads to the maximum pumped heat. We validated this model against a designed TE cooler. In this cooler, contrary to conventional TE coolers, where the temperature of the cold space is generally controlled at a specific temperature, and the performance of the cooler overlooked, the entropy generation and heat loss are engineered, and the electrical current is tuned to minimize {\gamma} by the controller so that the TE cooler works near to its optimum performance at any time., Comment: 14 pages, 7 figures

We explore the application of an extrapolative method that yields very accurate total and relative energies from variational and diffusion quantum Monte Carlo (VMC and DMC) results. For a trial wave function consisting of a small configuration interaction (CI) wave function obtained from full CI quantum Monte Carlo and reoptimized in the presence of a Jastrow factor and an optional backflow transformation, we find that the VMC and DMC energies are smooth functions of the sum of the squared coefficients of the initial CI wave function, and that quadratic extrapolations of the non-backflow VMC and backflow DMC energies intersect within uncertainty of the exact total energy. With adequate statistical treatment of quasi-random fluctuations, the extrapolate and intersect with polynomials of order two (XSPOT) method is shown to yield results in agreement with benchmark-quality total and relative energies for the C2, N2, CO2, and H2O molecules, as well as for the C2 molecule in its first electronic singlet excited state, using only small CI expansion sizes., Comment: 10 pages, 8 figures (+ supplementary: 7 pages, 3 figures)

This paper explores the concept of \'{e}tal\'{e} spaces associated with residuated lattices. Notions of bundles and \'{e}tal\'{e}s of residuated lattices over a given topological space are introduced and investigated. For a topological space $\mathscr{B}$, we establish that the category of \'{e}tal\'{e}s of residuated lattices over $\mathscr{B}$ with morphisms of \'{e}tal\'{e}s of residuated lattices is coreflective in the category of bundles of residuated lattices over $\mathscr{B}$ along with morphisms of bundles of residuated lattices. We provide a method for transferring an \'{e}tal\'{e} of residuated lattices over a topological space to another, utilizing a continuous map. Finally, we define a contravariant functor, called the section functor, from the category of \'{e}tal\'{e}s of residuated lattices with inverse morphisms to the category of residuated lattices., Comment: 45 pages

The Y1-xLixBa2Cu3O7-d polycrystalline bulk superconductors doped with Li substituting at the Y site at different concentrations (x = 0, 0.01, 0.02, 0.1) were prepared using the citrate-gel method to study the effects of doping on the superconducting temperature and critical current density. The question was whether Li addition characterized by a high Debye frequency would have any positive effects on Tc. The optimum citrate-gel and heat treatment conditions were identified as those yielding samples with a maximum grain size on the order of 50 micro.m (up to the optimum Li-doping level, x=0.01). Li substitution at the Y site was verified by structural, electrical, and magnetic measurements of the produced samples, whereas X-ray diffraction (XRD) analysis revealed the formation of a pure phase with no visible impurity phases. Moreover, AC magnetic susceptibility measurements showed no increase in the superconducting transition temperature Tc, consistent with the predicted results obtained by machine learning method, although it was theoretically expected to increase owing to the high Debye frequency of Li. This observation is consistent with magnetic coupling models for pairing mechanism in cuprates. Finally, because of the optimum conditions of the preparation procedure, nearly identical values of the critical current density (Jc) were recorded for samples with different Li doping levels (up to the optimum Li doping level). It was found that improved compound preparation conditions would have a critical and extensive effect on Jc enhancement, with nearly no Tc suppression.

Gas turbine engines are highly efficient and powerful because of their high-pressure turbines (HPTs). Furthermore, stationary blades shape and prepare high-pressure gas for efficient utilization by moving blades. Consequently, optimizing the geometric features of both stationary and moving blades during the first and second stages of HPT is necessary. By considering stagger, inlet, and outlet angles of the first and second stages of blades as design variables and polytropic efficiency as an objective, this study examines HPT performance. The performance characteristics of the turbine are examined using Computational Fluid Dynamics (CFD). To model the objective functions of the design variables, the Design of Experiments (DOE) method is employed. A Genetic Algorithm (GA) optimizes torque, power, and polytropic efficiency. Optimization provides valuable insights into optimal design principles. As shown by the simulation results, stagger, inlet, and outlet angles affect turbine performance. Through GA optimization, torque, power, and polytropic efficiency are improved by 8.4%, 0.69%, and 1.2%, respectively. As a result of these improvements, the optimization approach has been demonstrated to be effective in optimizing turbine performance. Upon examining the recommended design points, it becomes clear that stagger, inlet, and outlet angles of blades have a greater impact on performance than others.

Nowadays, pulmonary Computed Tomography Angiography (CTA) is the main tool for detecting Pulmonary Embolism (PE). However, manual interpretation of CTA volume requires a radiologist, which is time-consuming and error-prone due to the specific conditions of lung tissue, large volume of data, lack of experience, and eye fatigue. Therefore, Computer-Aided Design (CAD) systems are used as a second opinion for the diagnosis of PE. The purpose of this article is to review, evaluate, and compare the performance of deep learning and traditional-based CAD system for diagnosis PE and to help physicians and researchers in this field. In this study, all articles available in databases such as IEEE, ScienceDirect, Wiley, Springer, Nature, and Wolters Kluwer in the field of PE diagnosis were examined using traditional and deep learning methods. From 2002 to 2023, 23 papers were studied to extract the articles with the considered limitations. Each paper presents an automatic PE detection system that we evaluate using criteria such as sensitivity, False Positives (FP), and the number of datasets. This research work includes recent studies, state-of-the-art research works, and a more comprehensive overview compared to previously published review articles in this research area., Comment: 22 pages, 6 figures, 5 tables

Often, recommendation systems employ continuous training, leading to a self-feedback loop bias in which the system becomes biased toward its previous recommendations. Recent studies have attempted to mitigate this bias by collecting small amounts of unbiased data. While these studies have successfully developed less biased models, they ignore the crucial fact that the recommendations generated by the model serve as the training data for subsequent training sessions. To address this issue, we propose a framework that learns an unbiased estimator using a small amount of uniformly collected data and focuses on generating improved training data for subsequent training iterations. To accomplish this, we view recommendation as a contextual multi-arm bandit problem and emphasize on exploring items that the model has a limited understanding of. We introduce a new offline sequential training schema that simulates real-world continuous training scenarios in recommendation systems, offering a more appropriate framework for studying self-feedback bias. We demonstrate the superiority of our model over state-of-the-art debiasing methods by conducting extensive experiments using the proposed training schema.

Over the last two decades the organic solar cell community has synthesised tens of thousands of novel polymers and small molecules in the search for an optimum light harvesting material. These materials were often crudely evaluated simply by measuring the current voltage curves in the light to obtain power conversion efficiencies (PCEs). Materials with low PCEs were quickly disregarded in the search for higher efficiencies. More complex measurements such as frequency/time domain characterisation that could explain why the material performed as it did were often not performed as they were too time consuming/complex. This limited feedback forced the field to advance using a more or less random walk of material development and has significantly slowed progress. Herein, we present a simple technique based on machine learning that can quickly and accurately extract recombination time constants and charge carrier mobilities as a function of light intensity simply from light/dark JV curves alone. This technique reduces the time to fully analyse a working cell from weeks to seconds and opens up the possibility of not only fully characterising new devices as they are fabricated, but also data mining historical data sets for promising materials the community has over looked., Comment: 10 figures, 30 pages

A numerical data set will be developed to assist in the design of Savonius wind turbines. The main objective of study is to improve Savonius turbine blade designs to increase torque coefficients, rotational speeds, and pressure coefficients. Simulating 3D models and validating them with wind tunnel data were part of the experimental design methodology. Multi-objective optimization is used to optimize turbine performance. Twist angle, aspect ratio, and overlap ratio are all important factors in determining the optimal torque and power coefficients. Data-driven objective functions were modeled using the group method of data handling (GMDH). Using an evolutionary Pareto-based optimization approach, polynomial models were used to plot Pareto fronts and TOPSIS to calculate optimal commercial points. The torque coefficient, rotational speed, and power coefficient are all improved by 13.74%, 0.071%, and 5.32%, respectively. As a result of the multi-objective optimization of the turbine, some significant characteristics of objective functions were discovered.

Asymptotic freedom is a feature of quantum chromodynamics that guarantees its well-posedeness. We derive an analog of asymptotic freedom enabling unconditional stability of lattice Boltzmann simulation of hydrodynamics. For the lattice Boltzmann models of nearly-incompressible flow, we show that the equilibrium based on entropy maximization is uniquely renormalizable. This results in a practical algorithm of constructing unconditionally stable lattice Boltzmann models.