Recursive Ground Truth Estimator for Social Data Streams

Yao, Shuochao; Amin, Tanvir Al; Lu, Shan; Hu, Shaohan; Li, S.; Wang, S.; Zhao, Yiran; Abdelzaher, Tarek; Kaplan, Lance; Aggarwal, Charų C.; Yener, Aylin

doi:10.1109/ipsn.2016.7460719

Cited by 29 publications

(10 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Wang et al [178] Wang et al [176] Wang et al [177] Wang et al [187] Wang et al [186] Huang and Wang [76] Marshall et al [109] Huang and Wang [77] Yao et al [201] MAP Models Ouyang et al [126] Ouyang et al [127] Yao et al [199] Optimization Methods Su et al [162] Meng et al [115] Ma et al [107] Others Miao et al [117] Qiu et al [133] Meng et al [116] some optimization-based approaches have also been presented, which find source reliability and entity truthfulness by assuming them to be constant in alternative iterations. Privacy-preserving truth discovery and predicting missing sensed information using matrix factorization are the other miscellaneous approaches available in the literature.…”

Section: Mle-based Emmentioning

confidence: 99%

“…Both models use the MAP estimator to find the optimal configuration of the random variables that maximize the posterior probability. Yao et al [199] identifies the inefficiency related to batch analysis of offline datasets and proposes an online recursive state estimator to recover ground truth from the streaming data. The sensing time is divided into equal-sized windows and posterior belief on source is computed at the end of each window by using the prior belief obtained at the beginning through MAP estimation.…”

Section: Binarymentioning

confidence: 99%

“…Both TSE and PTSE have been experimentally shown to take higher order polynomial time w.r.t inputs: (i) number of participants and (ii) number of events. Yao et al [199] Proposes an online recursive state estimator for recovering ground truth from streaming data. The objective is to speed up batch-like processing of sensory data generated by social sensing platforms.…”

Section: 33mentioning

confidence: 99%

“…By scalability, we imply that truth can be inferred in reasonable time event if the number of participants and the entities increase. Very few works in the existing literature have considered this issue; (6) With the exception of [200], EM algorithm-based works are mostly designed to run on static datasets. However, the computation becomes less efficient for streaming data as the iterative algorithm needs to re-run on the whole dataset from scratch every time it gets updated.…”

Section: How To Improve Existing Workmentioning

confidence: 99%

See 3 more Smart Citations

Quality of Information in Mobile Crowdsensing

Restuccia

Ghosh

Bhattacharjee

et al. 2017

ACM Trans. Sen. Netw.

152

View full text Add to dashboard Cite

Smartphones have become the most pervasive devices in people's lives, and are clearly transforming the way we live and perceive technology. Today's smartphones benefit from almost ubiquitous Internet connectivity and come equipped with a plethora of inexpensive yet powerful embedded sensors, such as accelerometer, gyroscope, microphone, and camera. This unique combination has enabled revolutionary applications based on the mobile crowdsensing paradigm, such as real-time road traffic monitoring, air and noise pollution, crime control, and wildlife monitoring, just to name a few. Differently from prior sensing paradigms, humans are now the primary actors of the sensing process, since they become fundamental in retrieving reliable and up-to-date information about the event being monitored. As humans may behave unreliably or maliciously, assessing and guaranteeing Quality of Information (QoI) becomes more important than ever. In this paper, we provide a new framework for defining and enforcing the QoI in mobile crowdsensing, and analyze in depth the current state-of-the-art on the topic. We also outline novel research challenges, along with possible directions of future work. F. Restuccia et al.Not only are today's smartphones ubiquitous devices, they are also equipped with a plethora of embedded multi-modal sensors; integrate wireless communication technologies such as 4G/WiFi Internet connectivity, and possess complex processing capabilities. For example, the cameras on smartphones can be used as video and image sensors [29], the microphone can be used as an acoustic sensor [39,154,206], and the embedded global positioning system (GPS) receiver can be used to gather accurate location information, while gyroscopes, accelerometers, and proximity sensors can be used to extract contextual information about the users, such as driving or walking states [75,122]. Further, additional sensors can be easily interfaced with smartphones via Bluetooth or wired connections, such as temperature, air quality, and humidity [69,168].These technological features, combined with the advanced sensing capability of humans, has spurred a significant amount of research from both academia and industry, which together have proposed over the last 10 years a myriad of applications based on the emerging mobile crowdsensing paradigm 1 . Mobile crowdsensing empowers ordinary citizens (or users 2 ) with the capability to actively monitor various phenomena pertaining to themselves (e.g., health, social connections) or their community (e.g., environment). This rich information or inference about themselves or the community may also be sent back to the participating or other concerned users to improve their life experiences, thus influencing their choices. Real-life applications, which can take advantage of both low-level sensor data and high-level user activities, range from real-time traffic monitoring applications, to environmental pollution monitoring, crime monitoring, and social networking, just to name a few. For a survey on mobile crowdsensin...

show abstract

Section: Mle-based Emmentioning

confidence: 99%

Section: Binarymentioning

confidence: 99%

Section: 33mentioning

confidence: 99%

Section: How To Improve Existing Workmentioning

confidence: 99%

See 2 more Smart Citations

Quality of Information in Mobile Crowdsensing

Restuccia

Ghosh

Bhattacharjee

et al. 2017

ACM Trans. Sen. Netw.

152

View full text Add to dashboard Cite

show abstract

“…Present-day sensing applications cover a broad range of areas including human interactions [27,59], context sensing [6,34,39,46,54], crowd sensing [55,58], object detection and tracking [7,29,42,53]. e recent commercial interest in IoT technologies promises a proliferation of smart objects in human spaces at a much broader scale.…”

Section: Introductionmentioning

confidence: 99%

DeepIoT

Yao

Zhao

Zhang

et al. 2017

Proceedings of the 15th ACM Conference on Embedded Network Sensor Systems

Self Cite

150

View full text Add to dashboard Cite

Recent advances in deep learning motivate the use of deep neutral networks in sensing applications, but their excessive resource needs on constrained embedded devices remain an important impediment. A recently explored solution space lies in compressing (approximating or simplifying) deep neural networks in some manner before use on the device. We propose a new compression solution, called DeepIoT, that makes two key contributions in that space. First, unlike current solutions geared for compressing speci c types of neural networks, DeepIoT presents a uni ed approach that compresses all commonly used deep learning structures for sensing applications, including fully-connected, convolutional, and recurrent neural networks, as well as their combinations. Second, unlike solutions that either sparsify weight matrices or assume linear structure within weight matrices, DeepIoT compresses neural network structures into smaller dense matrices by nding the minimum number of non-redundant hidden elements, such as lters and dimensions required by each layer, while keeping the performance of sensing applications the same. Importantly, it does so using an approach that obtains a global view of parameter redundancies, which is shown to produce superior compression. e compressed model generated by DeepIoT can directly use existing deep learning libraries that run on embedded and mobile systems without further modi cations. We conduct experiments with ve di erent sensing-related tasks on Intel Edison devices. DeepIoT outperforms all compared baseline algorithms with respect to execution time and energy consumption by a signi cant margin. It reduces the size of deep neural networks by 90% to 98.9%. It is thus able to shorten execution time by 71.4% to 94.5%, and decrease energy consumption by 72.2% to 95.7%. ese improvements are achieved without loss of accuracy. e results underscore the potential of DeepIoT for advancing the exploitation of deep neural networks on resource-constrained embedded devices.

show abstract