Dependable machine intelligence at the tactical edge

Archan Misra, Kasthuri Jayarajah, Dulanga Weerakoon, Randy Tandriansyah, Shuochao Yao, Tarek Abdelzaher

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

The paper describes a vision for dependable application of machine learning-based inferencing on resource-constrained edge devices. The high computational overhead of sophisticated deep learning learning techniques imposes a prohibitive overhead, both in terms of energy consumption and sustainable processing throughput, on such resource-constrained edge devices (e.g., audio or video sensors). To overcome these limitations, we propose a "cognitive edge" paradigm, whereby (a) an edge device first autonomously uses statistical analysis to identify potential collaborative IoT nodes, and (b) the IoT nodes then perform real-Time sharing of various intermediate state to improve their individual execution of machine intelligence tasks. We provide an example of such collaborative inferencing for an exemplar network of video sensors, showing how such collaboration can significantly improve accuracy, reduce latency and decrease communication bandwidth compared to non-collaborative baselines. We also identify various challenges in realizing such a cognitive edge, including the need to ensure that the inferencing tasks do not suffer catastrophically in the presence of malfunctioning peer devices. We then introduce the soon-To-be deployed Cognitive IoT testbed at SMU, explaining the various features that enable empirical testing of various novel edge-based ML algorithms.

Original languageEnglish (US)
Title of host publicationArtificial Intelligence and Machine Learning for Multi-Domain Operations Applications
EditorsTien Pham
PublisherSPIE
ISBN (Electronic)9781510626775
DOIs
StatePublished - Jan 1 2019
EventArtificial Intelligence and Machine Learning for Multi-Domain Operations Applications 2019 - Baltimore, United States
Duration: Apr 15 2019Apr 17 2019

Publication series

NameProceedings of SPIE - The International Society for Optical Engineering
Volume11006
ISSN (Print)0277-786X
ISSN (Electronic)1996-756X

Conference

ConferenceArtificial Intelligence and Machine Learning for Multi-Domain Operations Applications 2019
CountryUnited States
CityBaltimore
Period4/15/194/17/19

Fingerprint

intelligence
Sensors
Testbeds
Sensor
learning
Resources
self maneuvering units
Learning systems
resources
Statistical methods
Energy utilization
Vertex of a graph
time sharing
Throughput
Testbed
Statistical Analysis
Energy Consumption
Bandwidth
Latency
Baseline

Keywords

  • Cognitive Edge
  • Collaborative Sensing
  • Machine Learning
  • Tactical Edge
  • Video Analytics

ASJC Scopus subject areas

  • Electronic, Optical and Magnetic Materials
  • Condensed Matter Physics
  • Computer Science Applications
  • Applied Mathematics
  • Electrical and Electronic Engineering

Cite this

Misra, A., Jayarajah, K., Weerakoon, D., Tandriansyah, R., Yao, S., & Abdelzaher, T. (2019). Dependable machine intelligence at the tactical edge. In T. Pham (Ed.), Artificial Intelligence and Machine Learning for Multi-Domain Operations Applications [1100608] (Proceedings of SPIE - The International Society for Optical Engineering; Vol. 11006). SPIE. https://doi.org/10.1117/12.2522656

Dependable machine intelligence at the tactical edge. / Misra, Archan; Jayarajah, Kasthuri; Weerakoon, Dulanga; Tandriansyah, Randy; Yao, Shuochao; Abdelzaher, Tarek.

Artificial Intelligence and Machine Learning for Multi-Domain Operations Applications. ed. / Tien Pham. SPIE, 2019. 1100608 (Proceedings of SPIE - The International Society for Optical Engineering; Vol. 11006).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Misra, A, Jayarajah, K, Weerakoon, D, Tandriansyah, R, Yao, S & Abdelzaher, T 2019, Dependable machine intelligence at the tactical edge. in T Pham (ed.), Artificial Intelligence and Machine Learning for Multi-Domain Operations Applications., 1100608, Proceedings of SPIE - The International Society for Optical Engineering, vol. 11006, SPIE, Artificial Intelligence and Machine Learning for Multi-Domain Operations Applications 2019, Baltimore, United States, 4/15/19. https://doi.org/10.1117/12.2522656
Misra A, Jayarajah K, Weerakoon D, Tandriansyah R, Yao S, Abdelzaher T. Dependable machine intelligence at the tactical edge. In Pham T, editor, Artificial Intelligence and Machine Learning for Multi-Domain Operations Applications. SPIE. 2019. 1100608. (Proceedings of SPIE - The International Society for Optical Engineering). https://doi.org/10.1117/12.2522656
Misra, Archan ; Jayarajah, Kasthuri ; Weerakoon, Dulanga ; Tandriansyah, Randy ; Yao, Shuochao ; Abdelzaher, Tarek. / Dependable machine intelligence at the tactical edge. Artificial Intelligence and Machine Learning for Multi-Domain Operations Applications. editor / Tien Pham. SPIE, 2019. (Proceedings of SPIE - The International Society for Optical Engineering).
@inproceedings{e750925137e5469e925bc61c7e0c6d0b,
title = "Dependable machine intelligence at the tactical edge",
abstract = "The paper describes a vision for dependable application of machine learning-based inferencing on resource-constrained edge devices. The high computational overhead of sophisticated deep learning learning techniques imposes a prohibitive overhead, both in terms of energy consumption and sustainable processing throughput, on such resource-constrained edge devices (e.g., audio or video sensors). To overcome these limitations, we propose a {"}cognitive edge{"} paradigm, whereby (a) an edge device first autonomously uses statistical analysis to identify potential collaborative IoT nodes, and (b) the IoT nodes then perform real-Time sharing of various intermediate state to improve their individual execution of machine intelligence tasks. We provide an example of such collaborative inferencing for an exemplar network of video sensors, showing how such collaboration can significantly improve accuracy, reduce latency and decrease communication bandwidth compared to non-collaborative baselines. We also identify various challenges in realizing such a cognitive edge, including the need to ensure that the inferencing tasks do not suffer catastrophically in the presence of malfunctioning peer devices. We then introduce the soon-To-be deployed Cognitive IoT testbed at SMU, explaining the various features that enable empirical testing of various novel edge-based ML algorithms.",
keywords = "Cognitive Edge, Collaborative Sensing, Machine Learning, Tactical Edge, Video Analytics",
author = "Archan Misra and Kasthuri Jayarajah and Dulanga Weerakoon and Randy Tandriansyah and Shuochao Yao and Tarek Abdelzaher",
year = "2019",
month = "1",
day = "1",
doi = "10.1117/12.2522656",
language = "English (US)",
series = "Proceedings of SPIE - The International Society for Optical Engineering",
publisher = "SPIE",
editor = "Tien Pham",
booktitle = "Artificial Intelligence and Machine Learning for Multi-Domain Operations Applications",

}

TY - GEN

T1 - Dependable machine intelligence at the tactical edge

AU - Misra, Archan

AU - Jayarajah, Kasthuri

AU - Weerakoon, Dulanga

AU - Tandriansyah, Randy

AU - Yao, Shuochao

AU - Abdelzaher, Tarek

PY - 2019/1/1

Y1 - 2019/1/1

N2 - The paper describes a vision for dependable application of machine learning-based inferencing on resource-constrained edge devices. The high computational overhead of sophisticated deep learning learning techniques imposes a prohibitive overhead, both in terms of energy consumption and sustainable processing throughput, on such resource-constrained edge devices (e.g., audio or video sensors). To overcome these limitations, we propose a "cognitive edge" paradigm, whereby (a) an edge device first autonomously uses statistical analysis to identify potential collaborative IoT nodes, and (b) the IoT nodes then perform real-Time sharing of various intermediate state to improve their individual execution of machine intelligence tasks. We provide an example of such collaborative inferencing for an exemplar network of video sensors, showing how such collaboration can significantly improve accuracy, reduce latency and decrease communication bandwidth compared to non-collaborative baselines. We also identify various challenges in realizing such a cognitive edge, including the need to ensure that the inferencing tasks do not suffer catastrophically in the presence of malfunctioning peer devices. We then introduce the soon-To-be deployed Cognitive IoT testbed at SMU, explaining the various features that enable empirical testing of various novel edge-based ML algorithms.

AB - The paper describes a vision for dependable application of machine learning-based inferencing on resource-constrained edge devices. The high computational overhead of sophisticated deep learning learning techniques imposes a prohibitive overhead, both in terms of energy consumption and sustainable processing throughput, on such resource-constrained edge devices (e.g., audio or video sensors). To overcome these limitations, we propose a "cognitive edge" paradigm, whereby (a) an edge device first autonomously uses statistical analysis to identify potential collaborative IoT nodes, and (b) the IoT nodes then perform real-Time sharing of various intermediate state to improve their individual execution of machine intelligence tasks. We provide an example of such collaborative inferencing for an exemplar network of video sensors, showing how such collaboration can significantly improve accuracy, reduce latency and decrease communication bandwidth compared to non-collaborative baselines. We also identify various challenges in realizing such a cognitive edge, including the need to ensure that the inferencing tasks do not suffer catastrophically in the presence of malfunctioning peer devices. We then introduce the soon-To-be deployed Cognitive IoT testbed at SMU, explaining the various features that enable empirical testing of various novel edge-based ML algorithms.

KW - Cognitive Edge

KW - Collaborative Sensing

KW - Machine Learning

KW - Tactical Edge

KW - Video Analytics

UR - http://www.scopus.com/inward/record.url?scp=85072567887&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85072567887&partnerID=8YFLogxK

U2 - 10.1117/12.2522656

DO - 10.1117/12.2522656

M3 - Conference contribution

AN - SCOPUS:85072567887

T3 - Proceedings of SPIE - The International Society for Optical Engineering

BT - Artificial Intelligence and Machine Learning for Multi-Domain Operations Applications

A2 - Pham, Tien

PB - SPIE

ER -