jams_batch2024090309134020240903091340Scifiniticontact@jams.pubScifinitiComputing&AI ConnectCAIC3006-41636202411HieuLeJianTaoLossy compression has become essential an important technique to reduce data size in many domains. This type of compression is especially valuable for large-scale scientific data, whose size ranges up to several petabytes. Although Autoencoder-based models have been successfully leveraged to compress images and videos, such neural networks have not widely gained attention in the scientific data domain. Our work presents a neural network that not only significantly compresses large-scale scientific data, but also maintains high reconstruction quality. The proposed model is tested with scientific benchmark data available publicly and applied to a large-scale high-resolution climate modeling data set. Our model achieves a compression ratio of 140 on several benchmark data sets without compromising the reconstruction quality. 2D simulation data from the High-Resolution Community Earth System Model (CESM) Version 1.3 over 500 years are also being compressed with a compression ratio of 200 while the reconstruction error is negligible for scientific analysis.61320241CAIC.2024.193132https://creativecommons.org/licenses/by/4.0/10.69709/CAIC.2024.193132https:///article/1/1/11https:///article/1/1/11/pdfSLAC National Accelerator Laboratory. 2023. Linac coherent light source (lcls-ii) [Online]https://lcls.slac.stanford.edu/The National Radio Astronomy Observatory. 2023. The very large array radio telescope [Online]https://public.nrao.edu/10.1029/2020MS002298SonData compression for the exascale computing era-surveySupercomputing frontiers and innovations20141276DeutschGzip file format specification version 4.3.3Tech. Rep.199610.1145/103085.103089KimQmcpack: an open source ab initio quantum monte carlo package for the electronic structure of atoms, molecules and solidsJournal of Physics: Condensed Matter20183019195901VaswaniAttention is all you needAdvances in Neural Information Processing Systems201730DevlinBert: Pre-training of deep bidirectional transformers for language understandingarXiv preprint2018JohnstonComputationally efficient neural image compressionarXiv preprint2019KingmaAuto-encoding variational bayesarXiv preprint2013VahdatNvae: A deep hierarchical variational autoencoderAdvances in Neural Information Processing Systems2020331966710.1109/TBDATA.2021.306615110.1109/TVCG.2014.234645810.1109/TVCG.2006.143BalléEnd-to-end optimized image compressionarXiv preprint2016BalléVariational image compression with a scale hyperpriorarXiv preprint201810.1147/rd.203.019810.1147/rd.232.014910.1109/JRPROC.1952.273898MinnenJoint autoregressive and hierarchical priors for learned image compressionAdvances in Neural Information Processing Systems2018312023. Bpg image format [Online]https://bellard.org/bpg/10.1145/3641817Van Den OordNeural discrete representation learningAdvances in neural information processing systems20173010.1145/342262210.1145/3450626.345978510.1103/PhysRevFluids.5.11460210.1016/j.jcp.2018.10.04510.1039/D2LC00303A10.5194/nhess-23-1755-202310.1007/s10915-022-02059-4HendrycksGaussian error linear units (gelus)arXiv preprint2016Ballé, Simoncelli, EP. Density modeling of images using a generalized normalization transformation. arXiv preprint2015SønderbyLadder variational autoencodersAdvances in Neural Information Processing Systems201629RazaviGenerating diverse high-fidelity images with vq-vae-2Advances in Neural Information Processing Systems201932BengioEstimating or propagating gradients through stochastic neurons for conditional computationarXiv preprint201310.1016/j.jnca.2022.103452MariiaBaidachnaHaneenFatimaRahafOmranNourGhadbanMuhammadImranAhmadTahaLinaMohjaziThe traditional evaluation of smile esthetics is limited by subjectivity and latency, and the performance of machine learning methods within the dental sector is bounded by the lack of publicly available datasets. This paper presents a framework for a smart diagnostic tool for dental smile analysis. To accurately and efficiently identify esthetic issues from a single image of a smile, a convolutional neural network (CNN) and a segmentation model were trained. To mitigate overfitting, a diffusion model was employed to generate text-to-image data in addition to a manually curated dataset of real images. The CNN model was trained and evaluated on datasets containing different proportions of real and generated data, with the combined dataset model displaying an impressive accuracy of 81.61\% in detecting excessive gingival display. In order to identify several esthetic issues from one image, we examine instance segmentation as a possible smile arc classification technique. The proposed solution could be part of a standalone home-deployed smart mirror or connected to a network of an innovative Internet-of-Mirrors to facilitate patient-dentist communication.12919701CAIC.2024.194138https://creativecommons.org/licenses/by/4.0/10.69709/CAIC.2024.194138https:///article/1/1/12https:///article/1/1/12/pdf10.1155/2020/8830200FedTrees: A Novel Computation-Communication Efficient Federated Learning Framework Investigated in Smart Gridshttps://arxiv.org/pdf/2210.0006010.1186/s12903-018-0673-510.1111/joor.12250LucaNothing to smile aboutNeuropsychiatric Disease and Treatment2014101999MaySmile AestheticsAesthetic Orthognathic Surgery and Rhinoplasty201925310.7759/cureus.3261210.1111/jopr.1235910.1038/s41405-020-0032-x10.1016/j.ajodo.2009.01.02110.2319/072813-562.110.17126/joralres.2020.059AgouComparison of digital and paper assessment of smile aesthetics perceptionJournal of International Society of Preventive & Community Dentistry202010565910.4103/2278-0203.173426PrasannaEvaluation of waiting period, recall period, and appointment scheduling of outpatients in a dental hospitalDrug Invention Today [Online]20191171580InglehartDo Waiting Times in Dental Offices Affect Patient Satisfaction and Evaluations of Patient-Provider Relationships? A Quasi-experimental StudyJournal of Dental Hygiene [Online]201690320310.1016/S1470-2045(23)00511-9DowneyWhat the COVID-19 Pandemic Has Exposed: The Findings of Five Global Health Workforce Professions [Online]2023SusiloUnderstanding digital technology trends in healthcare and preventive strategyInternational Journal of Health and Medical Sciences202143347ShilpaDigital healthcare: current trends, challenges and future perspectivesProceedings of the Future Technologies Conference (FTC) 20212022vol. 264510.7759/cureus.4122710.1109/MCE.2016.255687910.1016/j.cmpb.2020.10590310.1007/s11036-020-01700-610.1109/ACCESS.2021.3059858HeMask R-CNNarXiv preprint arXiv:1703.06870 [Online]201810.3390/jcm1103085210.1038/s41598-019-40414-y10.1016/j.imu.2021.10077910.1186/s40537-019-0197-010.1016/j.neunet.2021.07.01910.3390/informatics10010028ZhuData Augmentation in Emotion Classification Using Generative Adversarial NetworksarXiv preprint arXiv:1711.00648201710.1109/TETCI.2023.3298535Adobe Firefly [Online]https://firefly.adobe.com/GhaiTeledentistry during COVID-19 pandemicDiabetes & Metabolic Syndrome: Clinical Research & Reviews [Online]2020145933Vishnu VardhanBaligodugulaAshutoshGhimireFathiAmsaadNetwork segmentation is the practice of dividing a network into smaller subnetworks, or segments, primarily to enhance security. This technique helps establish different zones with varying security requirements, facilitating access control and data protection. By segregating the network, institutions can isolate critical resources, such as database servers, and restrict access to specific groups of users or devices. This paper introduces the Secure IIoT Network Segmentation Framework (SiNeSF), a generalized cybersecurity pattern adaptable to various network segmentation scenarios, particularly within Industrial Internet of Things (IIoT) environments. SiNeSF addresses the unique challenges posed by the proliferation of connected devices and the need for robust security measures. The framework provides a foundation for creating customized designs for IIoT network segmentation, offering comprehensive guidelines for establishing secure boundaries and defining policies for data flow between segments. Using SiNeSF, organizations can improve their security posture, minimize the risk of unauthorized access, and protect sensitive data in IIoT environments. This strategy supports the deployment of IIoT solutions while maintaining a high level of security and compliance with industry standards.82720241CAIC.2024.193182https://creativecommons.org/licenses/by/4.0/10.69709/CAIC.2024.193182https:///article/1/1/15https:///article/1/1/15/pdf10.1109/JIOT.2017.273490310.3390/s2317747010.3390/s22155836Ur RahmanSpectrum sensing using nmlmf algorithm in cognitive radio networks for health care monitoring applicationsInt. J. Adv. Trends Comput. Sci. Eng.202095110.1109/MC.2011.29110.1109/JIOT.2020.3025916SinghShrew distributed denial-of-service (ddos) attack in iot applications: A surveyInternet of Things. Advances in Information and Communication Technology202497GhimirePower analysis side-channel attacks on same and cross-device settings: A survey of machine learning techniquesInternet of Things. Advances in Information and Communication Technology2024357Fabric defect segmentation system based on a lightweight gan for industrial internet of thingshttps://api.semanticscholar.org/CorpusID:24902432910.1109/TNSE.2022.318497510.1016/j.cose.2020.10216210.1016/j.comnet.2012.12.018BaligodugulaA comparative study of secure and efficient data duplication mechanisms for cloud-based iot applicationsProceedings of the 2023 International Conference on Advances in Computing Research (ACR’23)202356910.1109/TDSC.2019.295748810.1016/j.jnca.2023.103763AmjadMajidTomasVan RietbergenR VenkateshaPrasadEffective indoor navigation in the presence of dynamic obstacles is crucial for mobile robots. Previous research on deep reinforcement learning (DRL) for robot navigation has primarily focused on expanding neural network (NN) architectures and optimizing hardware setups. However, the impact of other critical factors, such as backward motion enablement, frame stacking buffer size, and the design of the behavioral reward function, on DRL-based navigation remains relatively unexplored. % To address this gap, we present a comprehensive analysis of these elements and their effects on the navigation capabilities of a DRL-controlled mobile robot. In our study, we developed a mobile robot platform and a ROS 2-based DRL navigation stack\footnote{The code is accessible via \url{https://github.com/amjadmajid/ROS2-for-DRL-autonomous-navigation-of-robots-with-LIDAR}}. Through extensive simulations and real-world experiments, we show the effects of said elements on the navigation of mobile robots. % Our findings reveal that our proposed agent achieves state-of-the-art performance in terms of navigation accuracy and efficiency. Notably, we identify the significance of backward motion enablement and a carefully designed behavioral reward function in enhancing the robot's navigation abilities. Finally, the insights gained from this research contribute to advancing the field of DRL-based robot navigation by uncovering the influence of crucial elements and providing valuable guidelines for designing robust navigation systems.80220241CAIC.2024.194188https://creativecommons.org/licenses/by/4.0/10.69709/CAIC.2024.194188https:///article/1/1/16https:///article/1/1/16/pdf10.1109/MRA.2006.163802210.1109/100.58097710.1109/SIECPC.2011.587694310.1109/CCNC49032.2021.936957210.1007/978-3-030-75472-310.26599/TST.2021.901001210.5555/3618408.3618475HiroseSelfi: Autonomous self-improvement with reinforcement learning for social navigationarXiv202410.1109/TITS.2023.331245310.1109/TIE.2023.329024410.1049/csy2.1202010.1145/3576842.358237310.1080/01691864.2017.1365009ParkHiql: Offline goal-conditioned rl with latent states as actionsAdv. Neural Inf. Process. Syst.20243610.1109/LRA.2021.309355110.1126/scirobotics.abm607410.1109/IROS.2004.138972710.1016/j.eswa.2016.06.02110.1109/IROS.2017.820213410.1109/ICRA.2019.8793979SurmannDeep reinforcement learning for real autonomous mobile robot navigation in indoor environmentsCoRR2020abs/2005.1385710.18178/ijmerr.9.8.1134-1139WeerakoonTERP: reliable planning in uneven outdoor environments using deep reinforcement learningCoRR2021abs/2109.0512010.1016/j.neunet.2020.12.00110.1016/j.knosys.2023.11135810.1109/SII.2017.827936010.3390/s20195493ParkTackling long-horizon tasks with model-based offline reinforcement learningarXiv2024LiuDecentralized structural-rnn for robot crowd navigation with deep reinforcement learningCoRR2020abs/2011.04820GaoA vision-based irregular obstacle avoidance framework via deep reinforcement learningCoRR2021abs/2108.0688710.1109/JSEN.2020.301629910.1038/s41586-024-07409-w10.48550/arXiv.2404.11884Constrained reinforcement learning for robotics via scenario-based programminghttps://arxiv.org/abs/2206.0960310.1126/scirobotics.adi756610.1109/SSCI47803.2020.9308468DudekComputational Principles of Mobile Robotics201010.1371/journal.pone.025275410.1109/CVPR.2018.0006110.1016/j.neucom.2020.07.061Spinning up in deep reinforcement learninghttps://spinningup.openai.com/en/latest/index.htmlBengioPractical recommendations for gradient-based training of deep architecturesCoRR2012abs/1206.5533MastersRevisiting small batch training for deep neural networksCoRR2018abs/1804.0761210.48550/arXiv.1509.02971FujimotoAddressing function approximation error in actor-critic methodsCoRR2018abs/1802.09477FanCrowdmove: Autonomous mapless navigation in crowded scenariosCoRR2018abs/1807.07870LongTowards optimally decentralized multi-robot collision avoidance via deep reinforcement learningCoRR2017abs/1709.10082MnihPlaying atari with deep reinforcement learningCoRR2013abs/1312.560210.1609/aaai.v32i1.1179610.3390/s20133664HouthooftCuriosity-driven exploration in deep reinforcement learning via bayesian neural networksCoRR2016abs/1605.0967410.1016/B978-1-55860-335-6.50030-1