375 Hits in 14.7 sec

One Size Does Not Fit All: Multi-Scale, Cascaded RNNs for Radar Classification [article]

Dhrubojyoti Roy, Sangeeta Srivastava, Aditya Kusupati, Pranshu Jain, Manik Varma, Anish Arora
2019 arXiv   pre-print
lower tier, and a more complex RNN classifier for source classification at the upper tier.  ...  We find that this problem can be resolved by learning the classifier across multiple time-scales.  ...  Constrained Devices 2 .  ... 
arXiv:1909.03082v1 fatcat:wl277f3f3ra5tkvnbj322d7bqi

A Unifying Review of Deep and Shallow Anomaly Detection

Lukas Ruff, Jacob R. Kauffmann, Robert A. Vandermeulen, Gregoire Montavon, Wojciech Samek, Marius Kloft, Thomas G. Dietterich, Klaus-Robert Muller
2021 Proceedings of the IEEE  
Furthermore, connections between classic "shallow" and novel deep approaches are established, and it is shown how this relation might cross-fertilize or extend both directions.  ...  ABSTRACT | Deep learning approaches to anomaly detection (AD) have recently improved the state of the art in detection performance on complex data sets, such as large Manuscript  ...  Further variants of deep one-class classification include multimodal [145] or time-series extensions [338] and methods that employ adversarial learning [138] , [141] , [339] or transfer learning  ... 
doi:10.1109/jproc.2021.3052449 fatcat:i65pl2azw5dv7mtq7w7q3ylxgq

Security Threats and Artificial Intelligence based Countermeasures for Internet of Things Networks: A Comprehensive Survey

Shakila Zaman, Khaled Alhazmi, Mohammed Aseeri, Muhammad Raisuddin Ahmed, Risala Tasin Khan, M. Shamim Kaiser, Mufti Mahmud
2021 IEEE Access  
RNN Works on graph-like structure to detect malicious data in time-series based threats DNN DNN processes the supplied data to recognize the pattern or predict the desired result more globally through  ...  [125] proposed a statistical anomaly-based attack detection system for auto-sustainable IoT devices using time-series analysis.  ... 
doi:10.1109/access.2021.3089681 fatcat:fatpywnjzzfilidakyduz6qz44

Deep Learning for IoT Big Data and Streaming Analytics: A Survey [article]

Mehdi Mohammadi, Ala Al-Fuqaha, Sameh Sorour, Mohsen Guizani
2018 arXiv   pre-print
Based on the nature of the application, these devices will result in big or fast/real-time data streams.  ...  In the era of the Internet of Things (IoT), an enormous amount of sensing devices collect and/or generate various sensory data over time for a wide range of fields and applications.  ...  constrained devices. 1) Network Compression: One way of adopting DNNs to resource-constrained devices is through the use of network compression, in which a dense network is converted to a sparse network  ... 
arXiv:1712.04301v2 fatcat:kr64lst37rhlfcpaxckgzlozvu

Deep learning for sensor-based activity recognition: A Survey

Jindong Wang, Yiqiang Chen, Shuji Hao, Xiaohui Peng, Lisha Hu
2018 Pattern Recognition Letters  
We also present detailed insights on existing work and propose grand challenges for future research.  ...  However, those methods often heavily rely on heuristic hand-crafted feature extraction, which could hinder their generalization performance.  ...  The main line of RNN based HAR models are dealing with resource-constrained environments while still achieve good performance.  ... 
doi:10.1016/j.patrec.2018.02.010 fatcat:xwtshq6ivnggjicqhn6ejzhksi

Deep Learning-based Cattle Activity Classification Using Joint Time-frequency Data Representation [article]

Seyedeh Faezeh Hosseini Noorbin, Siamak Layeghy, Brano Kusy, Raja Jurdak, Greg Bishop-hurley, Marius Portmann
2020 arXiv   pre-print
This shows the potential for our classification approach to run on resource-constrained embedded and IoT devices.  ...  classification accuracy.  ...  As a final caveat, the RNN-LSTM classifier is relatively resource intensive and requires a high memory bandwidth [38] , which makes it difficult to be deployed on resource-constrained edge devices.  ... 
arXiv:2011.03381v1 fatcat:hkjdu652tnecpael55yis4wmly

Artificial Neural Networks-Based Machine Learning for Wireless Networks: A Tutorial [article]

Mingzhe Chen, Ursula Challita, Walid Saad, Changchuan Yin, and Mérouane Debbah
2019 arXiv   pre-print
Next-generation wireless networks must support ultra-reliable, low-latency communication and intelligently manage a massive number of Internet of Things (IoT) devices in real-time, within a highly dynamic  ...  In a nutshell, this article constitutes one of the first holistic tutorials on the development of machine learning techniques tailored to the needs of future wireless networks.  ...  However, RNNs can require more time to train since each value of the activation function depends on the series data recorded in RNNs.  ... 
arXiv:1710.02913v2 fatcat:kljn2evlwba4fha4lpwxjpv4yu

Deep Architectures for Human Activity Recognition using Sensors

Zartasha Baloch, Faisal Karim Shaikh, Mukhtiar Ali Unar
2019 3C Tecnología  
The use of sensors in the field of HAR opens new avenues for machine learning (ML) researchers to accurately recognize human activities.  ...  In this paper, we have reviewed recent research studies on deep models for sensor-based human activity recognition. The aim of this article is to identify recent trends and challenges in HAR.  ...  RECURRENT NEURAL NETWORK Recurrent Neural Network (RNN) is a deep model with cyclic connections, which empowers it to capture correlations between time series data.  ... 
doi:10.17993/3ctecno.2019.specialissue2.14-35 fatcat:cinqa2t2uvd6bklp3vor6ljylm

Applying Deep Learning to Individual and Community Health Monitoring Data: A Survey

Zhen-Jie Yao, Jie Bi, Yi-Xin Chen
2018 International Journal of Automation and Computing  
We focus on 7 application areas of deep learning, which are electronic health records (EHR), electrocardiography (ECG), electroencephalogram (EEG), community healthcare, data from wearable devices, drug  ...  The performance is not very extraordinary, the main point is that the deep learning method is implemented on a very resource constrained device.  ...  [51] presented a human activity recognition technique based on deep learning methodology, which is designed to enable accurate and real-time classification for low-power wearable devices.  ... 
doi:10.1007/s11633-018-1136-9 fatcat:drp2ixw3dvb5thxxnuzl4vjqsu

KecNet: A Light Neural Network for Arrhythmia Classification Based on Knowledge Reinforcement

Peng Lu, Yang Gao, Hao Xi, Yabin Zhang, Chao Gao, Bing Zhou, Hongpo Zhang, Liwei Chen, Xiaobo Mao
2021 Journal of Healthcare Engineering  
Therefore, it is crucial to design deep learning (DL) algorithms that are more suitable for resource-constrained mobile devices.  ...  Acquiring electrocardiographic (ECG) signals and performing arrhythmia classification in mobile device scenarios have the advantages of short response time, almost no network bandwidth consumption, and  ...  and robust DL algorithm for resource-constrained mobile devices.  ... 
doi:10.1155/2021/6684954 pmid:33995984 pmcid:PMC8096590 fatcat:z5elmrmfu5behoqzo6xu5na5mi

Quantitative Analysis of Image Classification Techniques for Memory-Constrained Devices [article]

Sebastian Müksch, Theo Olausson, John Wilhelm, Pavlos Andreadis
2020 arXiv   pre-print
This limits their usefulness in applications relying on embedded devices, where memory is often a scarce resource.  ...  Recently, there has been significant progress in the field of image classification on such memory-constrained devices, with novel contributions like the ProtoNN, Bonsai and FastGRNN algorithms.  ...  This limits the feasibility of applying CNNs to carry out image classification on memory-constrained devices.  ... 
arXiv:2005.04968v4 fatcat:5fnphvrpr5hg5jwjiau7rzdsne

Energy time series forecasting-analytical and empirical assessment of conventional and machine learning models

Hala Hamdoun, Alaa Sagheer, Hassan Youness
2021 Journal of Intelligent & Fuzzy Systems  
Machine learning methods have been adopted in the literature as contenders to conventional methods to solve the energy time series forecasting (TSF) problems.  ...  If a time series contains observations of one variable, it is denoted as a univariate time series (UTS), otherwise it is a multivariate time series (MTS) [87] .  ...  However, obtaining reasonably accurate predictions from the energy time series data is quite difficult due to many inevitable limitations.  ... 
doi:10.3233/jifs-201717 fatcat:jx37epearjeehkat32omjpjepa

WearableDL: Wearable Internet-of-Things and Deep Learning for Big Data Analytics—Concept, Literature, and Future

Aras R. Dargazany, Paolo Stegagno, Kunal Mankodiya
2018 Mobile Information Systems  
In recent times, wearable IoT devices have enabled the streaming of big data from smart wearables (e.g., smartphones, smartwatches, smart clothings, and personalized gadgets) to the cloud servers.  ...  In order to understand the current state-of-the-art, we systematically reviewed over 100 similar and recently published scientific works on the development of DL approaches for wearable and person-centered  ...  time-series signal classification.  ... 
doi:10.1155/2018/8125126 fatcat:ty3a7n4in5aahbqyl7wum5vonq

Self-Supervised Long-Short Term Memory Network for Solving Complex Job Shop Scheduling Problem

2021 KSII Transactions on Internet and Information Systems  
This paper proposes an effective scheduler based on deep learning technology named self-supervised long-short term memory (SS-LSTM) to handle complex JSSP accurately.  ...  Conventional methods are very time-consumption and cannot deal with complicated JSSP instances as it uses one optimal algorithm to solve JSSP.  ...  The learning-based methods mainly treated a JSSP as one sub-classification problem, consisting of shallow learning and deep learning methods.  ... 
doi:10.3837/tiis.2021.08.016 fatcat:tp2afe2h2vgxdmaufcwgeslz6m

Edge Intelligence: Paving the Last Mile of Artificial Intelligence With Edge Computing

Zhi Zhou, Xu Chen, En Li, Liekang Zeng, Ke Luo, Junshan Zhang
2019 Proceedings of the IEEE  
More recently, with the proliferation of mobile computing and Internet of Things (IoT), billions of mobile and IoT devices are connected to the Internet, generating zillions bytes of data at the network  ...  Finally, we discuss future research opportunities on EI. We believe that this survey will elicit escalating attentions, stimulate fruitful discussions, and inspire further research ideas on EI.  ...  The training of RNN is based on backpropagation through time (BPTT) [29] . Long short-term memory (LSTM) [30] is an extended version of RNNs.  ... 
doi:10.1109/jproc.2019.2918951 fatcat:d53vxmklgfazbmzjhsq3tuoama
« Previous Showing results 1 — 15 out of 375 results