Cloud-backed mobile cognition

Power-efficient deep learning in the autonomous vehicle era


Low-power embedded technology offers a roadmap for enabling deep learning (DL) applications in mobile scenarios, like future autonomous vehicles. However, the lack of breakthrough power efficiency improvements can jeopardize the realization of truly “cognitive” mobile systems that meet real-time deadlines. This work focuses on the new generation cloud-backed mobile cognition system architecture where vehicles execute DL applications with dynamic assistance from the cloud. We unveil opportunities for power-efficient inferencing at the edge through a technique that balances inference execution across the cloud and the vehicle. This level of adaptation results in significant power efficiency improvements compared to all or nothing solutions, where inferences execute either completely on the vehicle or completely in the cloud. In addition, the cloud can have an active role in helping the vehicle to improve its DL capabilities by communicating relevant model updates, with up to 63% bandwidth savings and negligible accuracy degradation when the proposed relevance-driven federated learning technique is used. Finally, the cloud-backed mobile cognition concept is extended to the case of “flying clouds” where vehicles connect to flying drones that provide services while in flight. Although their capabilities are not on par with the stationary cloud, the flying cloud reduces services’ latency significantly and enables critical functionalities.

This is a preview of subscription content, access via your institution.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11
Fig. 12
Fig. 13


  1. 1.

    The terms “cognitive computing” and AI are used interchangeably in this paper.

  2. 2.

    Peer-to-peer federated learning is also possible, where the aggregation is performed across the devices in a fully-distributed manner.


  1. 1.

    Callegaro D et al (2020) Dynamic distributed computing for infrastructure-assisted autonomous UAVs. ICC 2020:1–6

    Google Scholar 

  2. 2.

    Callegaro D, Levorato M (2018) Optimal computation offloading in edge-assisted UAV systems. GLOBECOM 2018:1–6

    Google Scholar 

  3. 3.

    Chen J, Ran X (2019) Deep learning with edge computing: A review. Proc IEEE 107:1655–1674

    Article  Google Scholar 

  4. 4.

    Chen Y et al (2016) DianNao family: Energy-efficient hardware accelerators for machine learning. Commun ACM 59:105–112

    Article  Google Scholar 

  5. 5.

    Cheng Y et al (2017) A survey of model compression and acceleration for deep neural networks. arXiv:1710.09282

  6. 6.

    Dolcourt J (2019) We Ran 5G Speed Tests on Verizon, AT&T, EE and More: Here’s What We Found.

  7. 7.

    Du Z et al (2015) ShiDianNao: Shifting vision processing closer to the sensor. ISCA 2015:92–104

    Article  Google Scholar 

  8. 8.

    Eliot L (2017) In-car voice commands NLP for self-driving cars.

  9. 9.

    Eshratifar AE, Pedram M (2018) Energy and performance efficient computation offloading for deep neural networks in a mobile cloud computing environment. GLSVLSI 2018:111–116

    Google Scholar 

  10. 10.

    Google LLC (2020) Edge TPU.

  11. 11.

    Han S et al (2015) Deep compression: Compressing deep neural networks with pruning, trained quantization and Huffman coding. arXiv:1510.00149

  12. 12.

    Howard A et al (2017) MobileNets: Efficient convolutional neural networks for mobile vision applications. arXiv:1704.04861

  13. 13.

    Iandola F et al (2016) SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and \(<\)0.5MB model size. arXiv:1602.07360

  14. 14.

    Kang Y et al (2017) Neurosurgeon: Collaborative intelligence between the cloud and mobile edge. ASPLOS 2017:615–629

    MathSciNet  Google Scholar 

  15. 15.

    Konečný J et al (2016) Federated learning: Strategies for improving communication efficiency. arXiv:1610.05492

  16. 16.

    Lane ND, Georgiev P (2015) Can deep learning revolutionize mobile sensing? HotMobile 2015:117–122

    Article  Google Scholar 

  17. 17.

    LeCun Y et al (1998) Gradient-based learning applied to document recognition. Proc IEEE 86–11:2278–2324

    Article  Google Scholar 

  18. 18.

    Li S et al (2017) FitCNN: A cloud-assisted lightweight convolutional neural network framework for mobile devices. RTCSA 2017:1–6

    Google Scholar 

  19. 19.

    Liu S et al (2017) Computer architectures for autonomous driving. IEEE Comput 50:18–25

    Article  Google Scholar 

  20. 20.

    Maas A et al (2011) Learning word vectors for sentiment analysis. ACL HLT 2011:142–150

    Google Scholar 

  21. 21.

    McMahan H et al (2016) Communication-efficient learning of deep networks from decentralized data. arXiv:1602.05629

  22. 22.

    Memeti S, Pllana S (2016) Combinatorial optimization of work distribution on heterogeneous systems. ICPPW 2016:151–160

    Google Scholar 

  23. 23.

    Newman D (2019) How AI is making sentiment analysis easy.

  24. 24.

    Ouarnoughi H et al (2019) Hierarchical platform for autonomous driving. INTESA 2019:7–12

    Article  Google Scholar 

  25. 25.

    Pakha C et al (2018) Reinventing video streaming for distributed vision analytics. HotCloud 2018:1

    Google Scholar 

  26. 26.

    Redmon J, Farhadi A (2017) YOLO9000: Better, faster, stronger. CVPR 2017:6517–6525

    Google Scholar 

  27. 27.

    Riley G, Henderson T (2010) The ns-3 network simulator. Springer, Berlin Heidelberg, pp 15–34

    Google Scholar 

  28. 28.

    Sandler M et al (2018) MobileNetV2: Inverted residuals and linear bottlenecks. CVPR 2018:4510–4520

    Google Scholar 

  29. 29.

    Sun K et al (2014) M2C: Energy efficient mobile cloud system for deep learning. INFOCOM 2014:167–168

    Google Scholar 

  30. 30.

    Sze V et al (2017) Efficient processing of deep neural networks: A tutorial and survey. Proc IEEE 105:2295–2329

    Article  Google Scholar 

  31. 31.

    The Wall Street Journal (2020) Alexa has a new skill: Asking when it doesn’t know.

  32. 32.

    Verhelst M, Moons B (2017) Embedded deep neural network processing: Algorithmic and processor techniques bring deep learning to IoT and edge devices. IEEE Solid-State Circuits Mag 9:55–65

    Article  Google Scholar 

  33. 33.

    Weston J et al (2015) Towards AI-complete question answering: A set of prerequisite toy tasks. arXiv:1502.05698

  34. 34.

    Zamani H et al (2020) Analyzing and learning from user interactions for search clarification. SIGIR 2020:1181–1190

    Article  Google Scholar 

  35. 35.

    Zhang C et al (2019) Deep learning in mobile and wireless networking: A survey. IEEE Commun Surv Tutorials 21:2224–2287

Download references

Author information



Corresponding author

Correspondence to Augusto Vega.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

This research was developed with funding from the Defense Advanced Research Projects Agency (DARPA). The views, opinions and/or other findings expressed are those of the authors and should not be interpreted as representing the official views or policies of the Department of Defense or the U.S. Government. This document is approved for public release: distribution unlimited.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Vega, A., Buyuktosunoglu, A., Callegaro, D. et al. Cloud-backed mobile cognition. Computing (2021).

Download citation


  • Distributed AI
  • Autonomous vehicles
  • Power efficiency

Mathematical Subject Code:

  • 68M14: Distributed systems