Truth Discovery for Spatio-Temporal Events from Crowdsourced Data - - PowerPoint PPT Presentation

▶
truth discovery for spatio temporal events from
SMART_READER_LITE
LIVE PREVIEW

Truth Discovery for Spatio-Temporal Events from Crowdsourced Data - - PowerPoint PPT Presentation

Dec 30, 2023 402 likes •667 views

Truth Discovery for Spatio-Temporal Events from Crowdsourced Data Daniel Garca Ulloa, Li Xiong, Vaidy Sunderam Emory University Logo Image Here Author (Institution) Short Title Date or Conference Content Introduction Truth

slide-1
SLIDE 1

Date or Conference Short Title Author (Institution)

Logo Image Here

Truth Discovery for Spatio-Temporal Events from Crowdsourced Data

Daniel García Ulloa, Li Xiong, Vaidy Sunderam

Emory University

slide-2
SLIDE 2

Date or Conference Short Title Author (Institution)

Logo Image Here

Content

  • Introduction
  • Truth Inference in Spatial Crowdsourcing

○ Graphical model ○ Bayesian estimation ○ Bayesian estimation and Kalman filter ○ Experimental Results

  • Conclusion
slide-3
SLIDE 3

Date or Conference Short Title Author (Institution)

Logo Image Here

Introduction

slide-4
SLIDE 4

Date or Conference Short Title Author (Institution)

Logo Image Here

Related Work

Truth Inference in Spatial Crowdsourcing

Based on the techniques used, the algorithms can be classified in: 1) Direct Computation. Do not model workers or tasks. E.g. Majority Voting, Median 2) Optimization

  • Develop a function that captures the relation between tasks and

workers.

  • Optimize

3) Iterative methods

slide-5
SLIDE 5

Date or Conference Short Title Author (Institution)

Logo Image Here

Related Work

Truth Inference in Spatial Crowdsourcing

Zheng, Yudian, et al. "Truth inference in crowdsourcing: is the problem solved?." Proceedings of the VLDB Endowment 10.5 (2017): 541-552.

slide-6
SLIDE 6

Date or Conference Short Title Author (Institution)

Logo Image Here

Truth Discovery

Truth Inference in Spatial Crowdsourcing

slide-7
SLIDE 7

Date or Conference Short Title Author (Institution)

Logo Image Here

Truth Discovery

Truth Inference in Spatial Crowdsourcing

slide-8
SLIDE 8

Date or Conference Short Title Author (Institution)

Logo Image Here

Related Work

  • Truth discovery methods in SC can be classified [12]

into:

○ Iterative (e.g. Truthfinder [7]) ○ Optimization-based (e.g. MalVoteCount [13]) ○ Probabilistic graphical models (e.g. Latent Truth Model [23])

although overlaps are possible

  • Several methods ([19,20,21]) do not consider mobile

data

  • Other methods (TSE [15]) do not consider events

changing over time

slide-9
SLIDE 9

Date or Conference Short Title Author (Institution)

Logo Image Here

Contributions

  • Present dynamic graphical model that describes

dependencies between hidden and observed variables.

  • Present a Bayesian estimation method based on this

model

  • Enhance the model by incorporating the event

model that explicitly describes correlations, and develop an algorithm that uses the Kalman filter to take advantage of the event model

  • Simulations and real-world data show the

performance (F1 measure) of the methods

slide-10
SLIDE 10

Date or Conference Short Title Author (Institution)

Logo Image Here

Truth Discovery

Graphical Model

slide-11
SLIDE 11

Date or Conference Short Title Author (Institution)

Logo Image Here

Truth Discovery

Modelling user reliability

Trustworthy Passive Aggressive Noisy Untrustworthy

(high TP and FP) (high TP,TN Low FP,FN) (low TP,TN high FP,FN) (high TN,FN)

slide-12
SLIDE 12

Date or Conference Short Title Author (Institution)

Logo Image Here

Truth Discovery

Bayesian Estimation

Estimated

slide-13
SLIDE 13

Date or Conference Short Title Author (Institution)

Logo Image Here

Truth Discovery

Bayesian Estimation and Kalman Filter algorithm

Recursively Update Z:

slide-14
SLIDE 14

Date or Conference Short Title Author (Institution)

Logo Image Here

Truth Discovery

Bayesian Estimation and Kalman Filter algorithm

Recursively Update :

slide-15
SLIDE 15

Date or Conference Short Title Author (Institution)

Logo Image Here

Truth Discovery

Bayesian Estimation and Kalman Filter algorithm

Recursively Update H:

slide-16
SLIDE 16

Date or Conference Short Title Author (Institution)

Logo Image Here

Truth Discovery

Bayesian Estimation and Kalman Filter algorithm

slide-17
SLIDE 17

Date or Conference Short Title Author (Institution)

Logo Image Here

Truth Discovery

Bayesian Estimation and Kalman Filter algorithm

Hidden Observed Estimated Predicted Updated

slide-18
SLIDE 18

Date or Conference Short Title Author (Institution)

Logo Image Here

Truth Discovery

Truth Inference Algorithm: Case Study Results

Potholes (blue squares) according to the government website and Waze reports of potholes (red dots) in Boston on 2/23/2015 from 12PM to 4PM Time series of potholes and waze reports. Each data point represents 4 hours.

2/22/15 - 2/23/15 - 2/24/15 - 2/24/15 - 2/25/15 - 2/25/15 - 2/26/15 -

slide-19
SLIDE 19

Date or Conference Short Title Author (Institution)

Logo Image Here

Truth Discovery

Truth Inference Algorithm: Case Study Results

slide-20
SLIDE 20

Date or Conference Short Title Author (Institution)

Logo Image Here

Truth Discovery

Truth Inference Algorithm: Simulation Results

slide-21
SLIDE 21

Date or Conference Short Title Author (Institution)

Logo Image Here

Truth Discovery

Truth Inference Algorithm: Running Times

slide-22
SLIDE 22

Date or Conference Short Title Author (Institution)

Logo Image Here

Conclusion

  • Method has wide variety of applications by

considering a default state

  • Can be generalized to more labels
  • Incorporating the event model allows the method to

depend less on the observed data and enhances the performance.

slide-23
SLIDE 23

Date or Conference Short Title Author (Institution)

Logo Image Here

References

slide-24
SLIDE 24

Date or Conference Short Title Author (Institution)

Logo Image Here

References

slide-25
SLIDE 25

Date or Conference Short Title Author (Institution)

Logo Image Here

The End

Thanks!