Advanced search in
Research products
arrow_drop_down
Searching FieldsTerms
Any field
arrow_drop_down
includes
arrow_drop_down
Include:
830 Research products, page 1 of 83

  • Publications
  • 2018-2022
  • Preprint
  • AE
  • English

10
arrow_drop_down
Date (most recent)
arrow_drop_down
  • Publication . Preprint . Article . 2022
    Open Access English
    Authors: 
    Lu Yu; Shichao Pei; Lizhong Ding; Jun Zhou; Longfei Li; Chuxu Zhang; Xiangliang Zhang;

    This paper studies learning node representations with graph neural networks (GNNs) for unsupervised scenario. Specifically, we derive a theoretical analysis and provide an empirical demonstration about the non-steady performance of GNNs over different graph datasets, when the supervision signals are not appropriately defined. The performance of GNNs depends on both the node feature smoothness and the locality of graph structure. To smooth the discrepancy of node proximity measured by graph topology and node feature, we proposed SAIL - a novel \underline{S}elf-\underline{A}ugmented graph contrast\underline{i}ve \underline{L}earning framework, with two complementary self-distilling regularization modules, \emph{i.e.}, intra- and inter-graph knowledge distillation. We demonstrate the competitive performance of SAIL on a variety of graph applications. Even with a single GNN layer, SAIL has consistently competitive or even better performance on various benchmark datasets, comparing with state-of-the-art baselines. Accepted by AAAI2022, 10 pages, 3 figures

  • Open Access English
    Authors: 
    Filipe Rodrigues; Nicola Ortelli; Michel Bierlaire; Francisco C. Pereira;
    Countries: Denmark, Switzerland

    Specifying utility functions is a key step towards applying the discrete choice framework for understanding the behaviour processes that govern user choices. However, identifying the utility function specifications that best model and explain the observed choices can be a very challenging and time-consuming task. This paper seeks to help modellers by leveraging the Bayesian framework and the concept of automatic relevance determination (ARD), in order to automatically determine an optimal utility function specification from an exponentially large set of possible specifications in a purely data-driven manner. Based on recent advances in approximate Bayesian inference, a doubly stochastic variational inference is developed, which allows the proposed DCM-ARD model to scale to very large and high-dimensional datasets. Using semi-artificial choice data, the proposed approach is shown to very accurately recover the true utility function specifications that govern the observed choices. Moreover, when applied to real choice data, DCM-ARD is shown to be able discover high quality specifications that can outperform previous ones from the literature according to multiple criteria, thereby demonstrating its practical applicability. 21 pages, 2 figures, 11 tables

  • Open Access English
    Authors: 
    Inaam Ilahi; Muhammad Usama; Junaid Qadir; Muhammad Umar Janjua; Ala Al-Fuqaha; Dinh Thai Huang; Dusit Niyato;
    Publisher: Institute of Electrical and Electronics Engineers (IEEE)
    Country: Australia

    Deep reinforcement learning (DRL) has numerous applications in the real world, thanks to its ability to achieve high performance in a range of environments with little manual oversight. Despite its great advantages, DRL is susceptible to adversarial attacks, which precludes its use in real-life critical systems and applications (e.g., smart grids, traffic controls, and autonomous vehicles) unless its vulnerabilities are addressed and mitigated. To address this problem, we provide a comprehensive survey that discusses emerging attacks on DRL-based systems and the potential countermeasures to defend against these attacks. We first review the fundamental background on DRL and present emerging adversarial attacks on machine learning techniques. We then investigate the vulnerabilities that an adversary can exploit to attack DRL along with state-of-the-art countermeasures to prevent such attacks. Finally, we highlight open issues and research challenges for developing solutions to deal with attacks on DRL-based intelligent systems.

  • Open Access English
    Authors: 
    Yanwu Yang; Bernard J. Jansen; Yinghui Yang; Xunhua Guo; Daniel Zeng;

    In sponsored search advertising, keywords serve as an essential bridge linking advertisers, search users and search engines. Advertisers have to deal with a series of keyword decisions throughout the entire lifecycle of search advertising campaigns. This paper proposes a multi-level and closed-form computational framework for keyword optimization (MKOF) to support various keyword decisions. Based on this framework, we develop corresponding optimization strategies for keyword targeting, keyword assignment and keyword grouping at different levels (e.g., market, campaign and adgroup). With two real-world datasets obtained from past search advertising campaigns, we conduct computational experiments to evaluate our keyword optimization framework and instantiated strategies. Experimental results show that our method can approach the optimal solution in a steady way, and it outperforms two baseline keyword strategies commonly used in practice. The proposed MKOF framework also provides a valid experimental environment to implement and assess various keyword strategies in sponsored search advertising. 21 pages, 3 figures,1 table

  • Open Access English
    Authors: 
    Ismail Shahin; Noor Ahmad Al Hindawi; Ali Bou Nassif; Adi Alhudhaif; Kemal Polat;

    Recent analysis on speech emotion recognition has made considerable advances with the use of MFCCs spectrogram features and the implementation of neural network approaches such as convolutional neural networks (CNNs). Capsule networks (CapsNet) have gained gratitude as alternatives to CNNs with their larger capacities for hierarchical representation. To address these issues, this research introduces a text-independent and speaker-independent SER novel architecture, where a dual-channel long short-term memory compressed-CapsNet (DC-LSTM COMP-CapsNet) algorithm is proposed based on the structural features of CapsNet. Our proposed novel classifier can ensure the energy efficiency of the model and adequate compression method in speech emotion recognition, which is not delivered through the original structure of a CapsNet. Moreover, the grid search approach is used to attain optimal solutions. Results witnessed an improved performance and reduction in the training and testing running time. The speech datasets used to evaluate our algorithm are: Arabic Emirati-accented corpus, English speech under simulated and actual stress corpus, English Ryerson audio-visual database of emotional speech and song corpus, and crowd-sourced emotional multimodal actors dataset. This work reveals that the optimum feature extraction method compared to other known methods is MFCCs delta-delta. Using the four datasets and the MFCCs delta-delta, DC-LSTM COMP-CapsNet surpasses all the state-of-the-art systems, classical classifiers, CNN, and the original CapsNet. Using the Arabic Emirati-accented corpus, our results demonstrate that the proposed work yields average emotion recognition accuracy of 89.3% compared to 84.7%, 82.2%, 69.8%, 69.2%, 53.8%, 42.6%, and 31.9% based on CapsNet, CNN, support vector machine, multi-layer perceptron, k-nearest neighbor, radial basis function, and naive Bayes, respectively. 19 pages, 11 figures

  • Publication . Article . Preprint . 2022
    Open Access English
    Authors: 
    Wayne Jordan Chetcuti; Tobias Haug; Leong Chuan Kwek; Luigi Amico;

    We study the persistent current in a system of SU($N$) fermions with repulsive interaction confined in a ring-shaped potential and pierced by an effective magnetic flux. By applying a combination of Bethe ansatz and numerical analysis, we demonstrate that, as a combined effect of spin correlations, interactions and applied flux a specific phenomenon can occur in the system: spinon creation in the ground state. As a consequence, peculiar features in the persistent current arise. The elementary flux quantum, which fixes the persistent current periodicity, is observed to evolve from a single particle one to an extreme case of fractional flux quantum, in which one quantum is shared by all the particles. We show that the persistent current depends on the number of spin components $N$, number of particles and interaction in a specific way that in certain physical regimes has universality traits. At integer filling fractions, the persistent current is suppressed above a threshold of the repulsive interaction by the Mott spectral gap. Despite its mesoscopic nature, the current displays a clear finite size scaling behavior. Specific parity effects in the persistent current landscape hold. 16 revtex pages, 11 figures. Accepted for Publication in SciPost Physics

  • Open Access English
    Authors: 
    Mohammed Jouhari; Abdulla Al-Ali; Emna Baccour; Amr Mohamed; Aiman Erbad; Mohsen Guizani; Mounir Hamdi;
    Country: Qatar

    Unmanned Aerial Vehicles (UAVs) have attracted great interest in the last few years owing to their ability to cover large areas and access difficult and hazardous target zones, which is not the case of traditional systems relying on direct observations obtained from fixed cameras and sensors. Furthermore, thanks to the advancements in computer vision and machine learning, UAVs are being adopted for a broad range of solutions and applications. However, Deep Neural Networks (DNNs) are progressing toward deeper and complex models that prevent them from being executed on-board. In this paper, we propose a DNN distribution methodology within UAVs to enable data classification in resource-constrained devices and avoid extra delays introduced by the server-based solutions due to data communication over air-to-ground links. The proposed method is formulated as an optimization problem that aims to minimize the latency between data collection and decision-making while considering the mobility model and the resource constraints of the UAVs as part of the air-to-air communication. We also introduce the mobility prediction to adapt our system to the dynamics of UAVs and the network variation. The simulation conducted to evaluate the performance and benchmark the proposed methods, namely Optimal UAV-based Layer Distribution (OULD) and OULD with Mobility Prediction (OULD-MP), were run in an HPC cluster. The obtained results show that our optimization solution outperforms the existing and heuristic-based approaches. Accepted in IEEE Internet of Things Journal

  • Open Access English
    Authors: 
    Simone Raponi; Savio Sciancalepore; Gabriele Oligeri; Roberto Di Pietro;
    Country: Netherlands

    Assisted-navigation applications have a relevant impact on our daily life. However, technological progress in virtualization technologies and Software-Defined Radios recently enabled new attack vectors, namely, road traffic poisoning. These attacks open up several dreadful scenarios, which are addressed in this contribution by identifying the associated challenges and proposing innovative countermeasures.

  • Open Access English
    Authors: 
    Waleed Alhazzani; Mohammed Alshahrani; Fayez Alshamsi; Ohoud Aljuhani; Khalid Eljaaly; Samaher Hashim; Rakan M. AlQahtani; Doaa Alsaleh; Zainab Al Duhailib; Haifa Algethamy; +41 more
    Publisher: Elsevier

    Abstract BackgroundThe rapid increase in coronavirus disease 2019 (COVID-19) cases during the subsequent waves in Saudi Arabia and other countries prompted the Saudi Critical Care Society (SCCS) to put together a panel of experts to issue evidence-based recommendations for the management of COVID-19 in the intensive care unit (ICU).MethodsThe SCCS COVID-19 panel included 51 experts with expertise in critical care, respirology, infectious disease, epidemiology, emergency medicine, clinical pharmacy, nursing, respiratory therapy, methodology, and health policy. All members completed an electronic conflict of interest disclosure form. The panel addressed 9 questions that are related to the therapy of COVID-19 in the ICU. We identified relevant systematic reviews and clinical trials, then used the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) approach as well as the evidence-to-decision framework (EtD) to assess the quality of evidence and generate recommendations.ResultsThe SCCS COVID-19 panel issued 12 recommendations on pharmacotherapeutic interventions (immunomodulators, antiviral agents, and anticoagulants) for severe and critical COVID-19, of which 3 were strong recommendations and 9 were weak recommendations. ConclusionThe SCCS COVID-19 panel used the GRADE approach to formulate recommendations on therapy for COVID-19 in the ICU. The EtD framework allows adaptation of these recommendations in different contexts. The SCCS guideline committee will update recommendations as new evidence becomes available.

  • Open Access English
    Authors: 
    J. G. O. Machado; Brian Hare; Olaf Scholten; Stijn Buitink; Arthur Corstanje; Heino Falcke; Jörg R. Hörandel; T. Huege; G. K. Krampah; Pragati Mitra; +8 more
    Publisher: American Geophysical Union
    Countries: Netherlands, Germany

    When a lightning flash is propagating in the atmosphere it is known that especially the negative leaders emit a large number of very high frequency (VHF) radio pulses. It is thought that this is due to streamer activity at the tip of the growing negative leader. In this work, we have investigated the dependence of the strength of this VHF emission on the altitude of such emission for two lightning flashes as observed by the Low Frequency ARray (LOFAR) radio telescope. We find for these two flashes that the extracted amplitude distributions are consistent with a power-law, and that the amplitude of the radio emissions decreases very strongly with source altitude, by more than a factor of 2 from 1 km altitude up to 5 km altitude. In addition, we do not find any dependence on the extracted power-law with altitude, and that the extracted power-law slope has an average around 3, for both flashes.

Send a message
How can we help?
We usually respond in a few hours.