CoChat
Sign Up
Cited by 157Privacy-Preserving Technologies in DataIoT and Edge/Fog ComputingDomain Adaptation and Few-Shot Learning

Applications of Federated Learning; Taxonomy, Challenges, and Research Trends

Open access
Shaheen, Momina, Farooq, Muhammad Shoaib, Umer, Tariq, Kim, Byung-Seo·2022·Electronics
DOI 10.3390/electronics11040670
Open this paper in CoChat — free accountOpen PDF

Abstract

The federated learning technique (FL) supports the collaborative training of machine learning and deep learning models for edge network optimization. Although a complex edge network with heterogeneous devices having different constraints can affect its performance, this leads to a problem in this area. Therefore, some research can be seen to design new frameworks and approaches to improve federated learning processes. The purpose of this study is to provide an overview of the FL technique and its applicability in different domains. The key focus of the paper is to produce a systematic literature review of recent research studies that clearly describes the adoption of FL in edge networks. The search procedure was performed from April 2020 to May 2021 with a total initial number of papers being 7546 published in the duration of 2016 to 2020. The systematic literature synthesizes and compares the algorithms, models, and frameworks of federated learning. Additionally, we have presented the scope of FL applications in different industries and domains. It has been revealed after careful investigation of studies that 25% of the studies used FL in IoT and edge-based applications and 30% of studies implement the FL concept in the health industry, 10% for NLP, 10% for autonomous vehicles, 10% for mobile services, 10% for recommender systems, and 5% for FinTech. A taxonomy is also proposed on implementing FL for edge networks in different domains. Moreover, another novelty of this paper is that datasets used for the implementation of FL are discussed in detail to provide the researchers an overview of the distributed datasets, which can be used for employing FL techniques. Lastly, this study discusses the current challenges of implementing the FL technique. We have found that the areas of medical AI, IoT, edge systems, and the autonomous industry can adapt the FL in many of its sub-domains; however, the challenges these domains can encounter are statistical heterogeneity, system heterogeneity, data imbalance, resource allocation, and privacy.

AI snapshot

Written from the paper’s full text.

TL;DR

This systematic literature review surveys federated learning applications, algorithms, datasets, and challenges across edge and industry domains, selecting 105 primary studies from an initial 7546 papers. It proposes a taxonomy of FL applications and identifies healthcare and IoT/edge as the most represented domains.

Key findings

  • From 7546 initial papers (2016–2020), 105 primary studies were selected; application shares were 30% healthcare, 25% IoT/edge, 10% NLP, 10% autonomous vehicles, 10% mobile services, 10% recommender systems, and 5% FinTech.
  • The review proposes a taxonomy for FL applications across domains and sub-domains, and compares algorithms/models including FedAvg, FedProx, FedMA, Scaffold, FedSVRG, and others.
  • Datasets for FL are discussed in detail, including LEAF, Federated EMNIST, Shakespeare, CIFAR-100, and StackOverflow, with Table 10 listing client/data-item counts such as Shakespeare (16,068 items, 715 clients) and StackOverflow (135,818,730 items, 342,477 clients).
  • Key challenges identified are statistical heterogeneity, system heterogeneity, data imbalance, resource allocation, and privacy; expensive communication is also discussed.

Methods

  • A systematic literature review (SLR) adapted from Kitchenham (2007), organized around research questions on FL applications, algorithms/tools, datasets, and challenges.
  • Search ran from April 2020 to May 2021 across ACM, IEEE, ScienceDirect, Springer, and Wiley; initial results were 7546 papers published 2016–2020, reduced to 478 by title/keyword, 185 by abstract, and 105 by full-text screening.
  • Data extraction used an Excel structure; the review synthesizes and compares algorithms, models, frameworks, datasets, and challenges, and proposes a taxonomy for FL in edge networks and other domains.

Limitations

  • The paper does not explicitly state limitations. Evident from the text: the search ended in May 2021 and the initial corpus covered 2016–2020, so later FL research is excluded.
  • Evident from the text: application percentages are derived from the 105 selected studies rather than an exhaustive census, and no formal quality appraisal beyond inclusion/exclusion criteria is described.
  • The paper’s future directions note that FedAvg convergence is application-dependent and that communication, computation, and storage cost optimization for edge devices needs further refinement; additional application areas remain unexplored.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.3390/electronics11040670

Cites

151 of 151 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 83 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    5% of the studies reviewed implement FL for FinTech.

    “It has been revealed after careful investigation of studies that 25% of the studies used FL in IoT and edge-based applications and 30% of studies implement the FL concept in the health industry, 10% for NLP, 10% for autonomous vehicles, 10% for mobile services, 10% for recommender systems, and 5% for FinTech.”
    Quote in the paperNot a substantial claim
  • Findingown data

    10% of the studies reviewed implement FL for mobile services.

    “It has been revealed after careful investigation of studies that 25% of the studies used FL in IoT and edge-based applications and 30% of studies implement the FL concept in the health industry, 10% for NLP, 10% for autonomous vehicles, 10% for mobile services, 10% for recommender systems, and 5% for FinTech.”
    Quote in the paperNot a substantial claim
  • Findingnone given

    The areas of medical AI, IoT, edge systems, and the autonomous industry can adapt FL in many of its sub-domains.

    “We have found that the areas of medical AI, IoT, edge systems, and the autonomous industry can adapt the FL in many of its sub-domains; however, the challenges these domains can encounter are statistical heterogeneity, system heterogeneity, data imbalance, resource allocation, and privacy.”
    Quote in the paperSubstantial claim
  • Findingown data

    30% of the studies reviewed implement the FL concept in the health industry.

    “It has been revealed after careful investigation of studies that 25% of the studies used FL in IoT and edge-based applications and 30% of studies implement the FL concept in the health industry, 10% for NLP, 10% for autonomous vehicles, 10% for mobile services, 10% for recommender systems, and 5% for FinTech.”
    Quote in the paperNot a substantial claim
  • Findingown data

    The number of published research articles and technical reports in the area of FL is increasing exponentially.

    “The number of published research articles and technical reports are increasing exponentially in the area of FL as depicted in Figure 6, thus making FL an emerging technique of machine learning.”
    Quote in the paperNot a substantial claim
  • Classified as probably not a substantial claim. 79
  • May combine two claims. 35

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    The most well-known federated learning initiatives are now being carried out on smartphones.

    “The most well-known federated learning initiatives are now being carried out on smartphones (as shown in Table 8); however, the same approaches can be used for PCs, IoT, and other edge devices such as autonomous vehicles.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    The initial search found 7546 papers published between 2016 and 2020.

    “Initially a total of 7546 papers published between 2016 and 2020 were found through the digital libraries”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingnone given

    Federated learning provides an extensive variety of possible applications in many areas such as NLP and IoT.

    “Federated learning provides an extensive variety of possible applications in many areas such as NLP, IoT, etc. (as shown in Figure 7).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Among the 105 primary studies, 8 were from 2016, 7 from 2017, 10 from 2018, 65 from 2019, and 35 from the first half of 2020.

    “From 105 research articles, eight were from the year 2016, followed by seven articles published in 2017, 10 from the year 2018, and 65 were published in the year 2019.Until the first half of 2020, there were 35 that were shortlisted for this study.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    478 papers were shortlisted after title and keyword filtering.

    “out of which 478 were shortlisted after going through the titles and keywords filter.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    185 publications remained after abstract review.

    “After reviewing the abstracts, 185 publications remained for the review conduction.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Findingnone given

    As of the time of the review, there is no systematic literature review with a discussion of datasets of FL and implementation techniques published.

    “To the best of our knowledge, there is no systematic literature review with a discussion of datasets of FL and implementation techniques published as of yet.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingnone given

    The challenges that medical AI, IoT, edge systems, and the autonomous industry can encounter in adopting FL are statistical heterogeneity, system heterogeneity, data imbalance, resource allocation, and privacy.

    “We have found that the areas of medical AI, IoT, edge systems, and the autonomous industry can adapt the FL in many of its sub-domains; however, the challenges these domains can encounter are statistical heterogeneity, system heterogeneity, data imbalance, resource allocation, and privacy.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingnone given

    For federated learning implementation, numerous datasets were accessible; some were public, while others were not.

    “For the federated learning implementation, there were numerous datasets accessible.Some were accessible to the public, while others were not, and this provides an overview of the publicly available datasets.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Findingnone given

    The devices may be unreliable, and an active gadget will frequently stop working after a certain iteration.

    “The devices may be unreliable, and an active gadget will frequently stop working after a certain iteration.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Findingnone given

    Federated networks are natively heterogeneous due to differences in network connectivity, hardware, power, communication, storage, and computing capacities of nodes.

    “The ferreted networks are natively heterogeneous due to differences in network connectivity (Wi-Fi, 3G, 4G, 5G), hardware (CPU, RAM), power (battery level), communication, storage, and computing capacities of nodes.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingcitation

    There are issues in FL such as resource allocation, data imbalance, and statistical heterogeneity.

    “There are some other issues of this field that can arise as challenging problems to be addressed such as resource allocation [183], data imbalance, statistical heterogeneity, etc., as depicted in Figure 9.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [183] Ren, J., Ni, W., Nie, G., and Tian, H. (2021). Research on Resource Allocation for Efficient Federated Learning. arXiv.

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    The survey reveals that healthcare and IoT have a vast implementation opportunity of FL models, as 30% and 25% of the selected studies used FL in healthcare and edge applications.

    “The survey reveals that healthcare and IoT have a vast implementation opportunity of FL models, as 30% and 25% of the selected studies used FL in healthcare and edge applications.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingreasoning

    FL imbalances can be divided into three categories: size imbalance, local imbalance, and global imbalance.

    “We divided these FL imbalances into three categories to make it easier to distinguish between them:”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Findingreasoning

    Due to device and network size-related limits, only a small fraction of end nodes is active at any given time.

    “Furthermore, due to device and network size-related limits, only a small fraction of end nodes is active at any given time.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingnone given

    RQ1 Applications related studies are [3,7,9,11,13,25-44].

    “01 | RQ1 | Applications | [3,7,9,11,13,25-44]”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingreasoning

    The applications of FL are cross-domain.

    “Thus, the applications of FL are cross-domain.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingown data

    The survey screened 7546 initial results, 478 title and keyword selected, 185 abstract selected, and 105 full text selected.

    “| Results | 7546 | 478 | 185 | 105 |”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    RQ2 Algorithms and models related studies are [3,45-60].

    “02 | RQ2 | Algorithms and models | [3,45-60]”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingnone given

    RQ3 Data sets related studies are [61-66].

    “03 | RQ3 | Data sets | [61-66]”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    RQ4 Challenges related studies are [3,5,12,13,19,25,26,30,45,46,49,50,67-73].

    “04 | RQ4 | Challenges | [3,5,12,13,19,25,26,30,45,46,49,50,67-73]”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    In Google GBoard, the algorithmic model is trained by the words typed by the user, and then the trained model is sent to the server.

    “This happened to be the first real-world application of FL where the algorithmic model is trained by the words typed by the user, and then the trained model is sent to the server.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [28] Ramaswamy, S., Mathews, R., Rao, K., and Beaufays, F. (2019). Federated Learning for Emoji Prediction in a Mobile Keyboard. arXiv.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationnone given

    FL is adopted particularly in circumstances where privacy concerns and the desire to develop algorithms collide.

    “FL is adopted particularly in circumstances where privacy concerns and the desire to develop algorithms collide.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationown data

    FL is an emerging technique of machine learning.

    “The number of published research articles and technical reports are increasing exponentially in the area of FL as depicted in Figure 6, thus making FL an emerging technique of machine learning.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Premisecitation

    Collaborative research among nations is significant when worldwide health emergencies, such as COVID-19, are being encountered.

    “Furthermore, collaborative research among nations is significant when worldwide health emergencies, such as COVID-19, are being encountered [80].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [80] (2020, June 14). Coronavirus (COVID-19) Events as They Happen. Available online: https://www.who.int/emergencies/diseases/novel-coronavirus-2019/events-as-they-happen.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Modern healthcare systems entail a collaboration among hospitals, research labs and institutes, and federal agencies for the betterment of healthcare nation-wide.

    “Modern healthcare systems entail a collaboration among hospitals, research labs and institutes, and federal agencies for the betterment of healthcare nation-wide [79].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [79] Friedman, 2010, Policy: Achieving a nationwide learning health system, Sci. Transl. Med. source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Owkin and Intel are researching how FL could be leveraged to protect patients' data privacy while also using the data for better diagnosis.

    “Healthcare Owkin [31] and Intel [32] are researching how FL could be leveraged to protect patients' data privacy while also using the data for better diagnosis.[7,79,[109][110][111][112][113]”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites [31] (2020, June 13). Federated Learning—OWKIN. Available online: https://owkin.com/federated-learning/.

      Not checked in the free run.

    • cites [32] (2020, June 13). Intel Works with University of Pennsylvania in Using Privacy-Preserving AI to Identify Brain Tumors|Intel Newsroom. Available online: https://newsroom.intel.com/news/intel-works-university-pennsylvania-using-privacy-preserving-ai-identify-brain-tumors/#gs.7wuma4.

      Not checked in the free run.

    • cites [7,79,[109][110][111][112][113]

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Interpretationreasoning

    FL may enable autonomous vehicles to behave more quickly and correctly, minimizing accidents and increasing safety.

    “Federated learning may enable autonomous vehicles to behave more quickly and correctly, minimizing accidents and increasing safety.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationnone given

    FL can potentially minimize latency.

    “Another important reason to use FL is that it can potentially minimize latency.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    FL is applied in open banking and in finance for anti-financial crime processes, loan risk prediction, and the detection of financial crimes.

    “Banking and finance The FL is applied in open banking and in finance for anti-financial crime processes, loan risk prediction, and the detection of financial crimes.[77,[118][119][120][121][122]”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [77,[118][119][120][121][122]

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    FL has two primary applications for self-driving automobiles: safeguarding user data privacy and reducing latency.

    “Federated learning has two primary applications for self-driving automobiles.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    The authors in [114] argued that federated learning may be utilized to tackle security concerns with the standard cloud-based method and eliminate such threats by limiting data transfer volume and speeding up the learning process of autonomous vehicles.

    “They argued that federated learning may be utilized to tackle this problem and eliminate such threats by limiting data transfer volume and speeding up the learning process of autonomous vehicles.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [114] Pokhrel, 2020, Federated Learning With Blockchain for Autonomous Vehicles: Analysis and Design Challenges, IEEE Trans. Commun. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    The researchers offered two approaches to reduce the cost of uplink transmission: structured updates and sketching updates.

    “The researchers offered two approaches to reduce the cost of uplink transmission in this study: structured updates and sketching updates.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [50] Konečný, J., McMahan, H.B., Yu, F.X., Richtárik, P., Suresh, A.T., and Bacon, D. (2016). Federated learning: Strategies for improving communication efficiency. arXiv.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    In [114], the authors claimed that due to a large number of self-driving cars on the road, it is necessary to respond quickly to real-world circumstances.

    “In [114], the authors claimed that due to a large number of self-driving cars on the road, it is necessary to respond quickly to real-world circumstances [114].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [114] Pokhrel, 2020, Federated Learning With Blockchain for Autonomous Vehicles: Analysis and Design Challenges, IEEE Trans. Commun. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Their DL model may be developed to be 99 percent as accurate as a model trained using conventional approaches, using a FL methodology.

    “Their DL model may be developed to be 99 percent as accurate as a model trained using conventional approaches, using a FL methodology; according to the cooperation.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [32] (2020, June 13). Intel Works with University of Pennsylvania in Using Privacy-Preserving AI to Identify Brain Tumors|Intel Newsroom. Available online: https://newsroom.intel.com/news/intel-works-university-pennsylvania-using-privacy-preserving-ai-identify-brain-tumors/#gs.7wuma4.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Premisecitation

    In the healthcare industry, data privacy and security are extremely difficult to manage.

    “In the healthcare industry, data privacy and security are extremely difficult to manage [123].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [123] Rieke, 2020, The future of digital health with federated learning, NPJ Digit. Med. source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    By merely apprising the algorithms with precise data rather than whole user information, federated learning could improve user privacy.

    “By merely apprising the algorithms with precise data rather than whole user information, federated learning could improve user privacy.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Konečný et al. (2016) set the algorithms through which each client can independently compute the update based on its local data to the current model.

    “Konečný et al. (2016) set the algorithms through which each client can independently compute the update based on its local data to the current model [50].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [50] Konečný, J., McMahan, H.B., Yu, F.X., Richtárik, P., Suresh, A.T., and Bacon, D. (2016). Federated learning: Strategies for improving communication efficiency. arXiv.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    XGBoost uses way fewer resources compared to other systems and scales far billions of examples.

    “Conclusively, XGBoost uses way fewer resources compared to other systems and scales far billions of examples [4].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [4] Ji, S., Saravirta, T., Pan, S., Long, G., and Walid, A. (2021). Emerging Trends in Federated Learning: From Model Fusion to Federated X Learning. arXiv.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Much research demonstrates that FL has the potential to revolutionize autonomous vehicles and the Internet of Things.

    “Much research demonstrate that FL has the potential to revolutionize autonomous vehicles and the Internet of Things [124].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [124] Elbir, A.M., Soner, B., and Coleri, S. (2020). Federated Learning for Vehicular Networks. arXiv.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Chen et al. described an end-to-end tree boosting system named XGBoost.

    “Chen et al. described an end-to-end tree boosting system named XGBoost [51].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [51] Chen, T., and Guestrin, C. (2016, January 13–17). XGBoost: A Scalable Tree Boosting System. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, CA, USA. source

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    FL has the potential to permit autonomous vehicles to respond more quickly and correctly, minimizing accidents and increasing safety.

    “Because conventional cloud-learning entails huge transfers of data and a slower learning rate, federated learning has the potential to permit autonomous vehicles to respond more quickly and correctly, minimizing accidents and increasing safety.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Premisenone given

    Conventional cloud-learning entails huge transfers of data and a slower learning rate.

    “Because conventional cloud-learning entails huge transfers of data and a slower learning rate, federated learning has the potential to permit autonomous vehicles to respond more quickly and correctly, minimizing accidents and increasing safety.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Basnayake developed a method to improve sensor measurement reliabilities in a mobile robot network.

    “Basnayake, V. developed a method to improve sensor measurement reliabilities in a mobile robot network [180].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [180] Basnayake, V. (2019). Federated Learning for Enhanced Sensor Realiabity of Automated Wireless Networks. [Master’s Thesis, University of Oulu].

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisenone given

    CFL handles diverse client populations over time and is flexible enough to preserve privacy.

    “CFL handles diverse client populations over time and is also flexible enough to preserve privacy.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [169] Sattler, 2020, Clustered Federated Learning : Model-Agnostic Distributed Multi-Task Optimization under Privacy Constraints, IEEE Trans. Neural Netw. Learn. Syst. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Xu et al. performed a survey focusing on the progress of federated learning in healthcare informatics.

    “Xu et al. performed a survey focusing on the progress of federated learning in healthcare informatics [7].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [7] Xu, 2020, Federated Learning for Healthcare Informatics, J. Healthc. Inform. Res. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    The three algorithms were evaluated on MINIST dataset with non-i.i.d. and i.i.d. partitioning of data.

    “These algorithms were evaluated on MINIST dataset with the use of non-i.i.d. and i.i.d.partitioning of data.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [125] Nilsson, A., Smith, S., Ulm, G., Gustavsson, E., and Jirstrand, M. (2018, January 10–11). A Performance Evaluation of Federated Learning Algorithms. Proceedings of the Second Workshop on Distributed Infrastructures for Deep Learning, Rennes, France. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Basnayake used federated learning to detect sensor failures and estimate sensor properties in a distributed manner.

    “To detect sensor failures and estimate sensor properties in a distributed manner, they used federated learning.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [180] Basnayake, V. (2019). Federated Learning for Enhanced Sensor Realiabity of Automated Wireless Networks. [Master’s Thesis, University of Oulu].

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Chawla et al. proposed an over-sampling technique named SMOTE, which generates minority classes records to rebalance the data sample.

    “Chawla et al. proposed an over-sampling technique named SMOTE (synthetic minority over-sampling technique), which generates minority classes records to rebalance the data sample [126].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [126] Chawla, 2002, SMOTE: Synthetic Minority Over-sampling Technique, J. Artif. Intell. Res. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Yang et al. proposed data networks building among organizations to share data without compromising the privacy of the user.

    “Besides, based on federated mechanisms, they proposed data networks building among organizations to share data without compromising the privacy of the user [9].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [9] Samarakoon, S., Bennis, M., Saad, W., and Debbah, M. (2018, January 9–13). Federated Learning for Ultra-Reliable Low-Latency V2V Communications. Proceedings of the 2018 IEEE Global Communications Conference (GLOBECOM), Abu Dhabi, United Arab Emirates. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    The In-Edge AI framework was demonstrated to reach near-optimal performance with relatively low training overhead.

    “While this framework was demonstrated to reach near-optimal performance, the overhead of training was relatively low.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [179] Wang, 2018, Removing Confounding Factors Associated Weights in Deep Neural Networks Improves the Prediction Accuracy for Healthcare Applications, Biocomputing

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    XGBoost and AdaBoost can reduce bias as they learn from misclassification, but they are subtle to outliers and noise.

    “Some other approaches, such as Xgboost [51] and AdaBoost, can reduce bias as it learns from misclassification.However, these algorithms are subtle to outliers and noise.[178]”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [51] Chen, T., and Guestrin, C. (2016, January 13–17). XGBoost: A Scalable Tree Boosting System. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, CA, USA. source

      Not checked in the free run.

    • cites [178] Liang, X., Liu, Y., Chen, T., Liu, M., and Yang, Q. (2019). Federated Transfer Reinforcement Learning for Autonomous Driving. arXiv.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Wang et al. proposed the integration of DRL and FL to improve edge systems.

    “Wang et al. proposed the integration of deep reinforcement learning (DRL) as well as federated learning to improve edge systems [179].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [179] Wang, 2018, Removing Confounding Factors Associated Weights in Deep Neural Networks Improves the Prediction Accuracy for Healthcare Applications, Biocomputing

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    CFL applies to deep neural networks and gives strong mathematical guarantees on clustering quality.

    “It applies to deep neural networks, and, on the clustering quality, it gives strong mathematical guarantees.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [169] Sattler, 2020, Clustered Federated Learning : Model-Agnostic Distributed Multi-Task Optimization under Privacy Constraints, IEEE Trans. Neural Netw. Learn. Syst. source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Yang et al. proposed frameworks for secure federated learning, including horizontal and vertical FL as well as federated transfer learning.

    “Yang et al. proposed frameworks for secure federated learning [11].They introduced a comprehensive secure FL framework, including horizontal and vertical FL as well as federated transfer learning.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [11] Yang, 2019, Federated Machine Learning : Concept and Applications, ACM Trans. Intell. Syst. Technol. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    In [181], the authors introduced a generalization of Dropout, the DropConnect, to regularize the large and fully connected layers in neural networks.

    “In [181], the authors introduced a generalization of Dropout, the DropConnect to regularize the large and fully connected layers in neural networks.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [181] Zeiler, 2012, Regularization of Neural Networks using DropConnect, Int. Conf. Mach. Learn.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Mohri et al. proved convergence bounds supposing a convex hypothesis set and a convex loss function.

    “For this, they also proved convergence bounds supposing a convex hypothesis set, as well as a convex loss function.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [147] Mohri, M., Sivek, G., and Suresh, A.T. (2019, January 9–15). Agnostic Federated Learning. Proceedings of the 36th International Conference on Machine Learning, Long Beach, CA, USA.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Aggregated multiple DropConnect trained models showed state-of-the-art results on different image recognition benchmarks.

    “They aggregated multiple DropConnect trained models, and, on different image recognition benchmarks, they showed state-of-theart results.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [181] Zeiler, 2012, Regularization of Neural Networks using DropConnect, Int. Conf. Mach. Learn.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Mohri et al. optimized a centralized model in a newly proposed framework of agnostic federated learning.

    “Mohri et al. optimized a centralized model in a newly proposed framework of agnostic federated learning [147].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [147] Mohri, M., Sivek, G., and Suresh, A.T. (2019, January 9–15). Agnostic Federated Learning. Proceedings of the 36th International Conference on Machine Learning, Long Beach, CA, USA.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    The synthetic minority over-sampling technique (SMOTE) is one of the oversampling techniques used to address imbalance.

    “To address this issue, the synthetic minority over-sampling technique (SMOTE) is one of the oversampling tech-niques used.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [127] Han, 2005, Borderline-SMOTE : A New Over-Sampling Method in Imbalanced Data Sets Learning, Advances in Intelligent Computing source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Premisenone given

    Different client devices deconstruct the federated datasets.

    “Different client devices deconstruct the federated datasets.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Two new minorities were proposed by this method using borderline-SMOTE 1 and borderline-SMOTE 2 over the sampling method.

    “Two new minorities were proposed by this method using borderline-SMOTE 1 and borderline-SMOTE 2 over the sampling method.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [127] Han, 2005, Borderline-SMOTE : A New Over-Sampling Method in Imbalanced Data Sets Learning, Advances in Intelligent Computing source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    DropConnect sets a subset of weights in the system to zero, in contrast to Dropout which sets randomly selected activation subsets to zero.

    “In contrast to Dropout, which sets the randomly selected activation subsets to zero in each segment, DropConnect sets a subset of weights in the system to zero.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [181] Zeiler, 2012, Regularization of Neural Networks using DropConnect, Int. Conf. Mach. Learn.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Mohri et al. suggested that the framework yields a notion of fairness.

    “They suggested that the framework yields a notion of fairness.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [147] Mohri, M., Sivek, G., and Suresh, A.T. (2019, January 9–15). Agnostic Federated Learning. Proceedings of the 36th International Conference on Machine Learning, Long Beach, CA, USA.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Premisenone given

    CFL is considered a post-processing mechanism and is achieving similar or more significant goals than FL.

    “CFL is considered a post-processing mechanism and is achieving similar or more significant goals than the FL.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [169] Sattler, 2020, Clustered Federated Learning : Model-Agnostic Distributed Multi-Task Optimization under Privacy Constraints, IEEE Trans. Neural Netw. Learn. Syst. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Han et al. introduced the value of imbalanced datasets as well as their broad application areas in data mining.

    “Han et al. introduced the value of imbalanced datasets as well as their broad application areas in data mining [127].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [127] Han, 2005, Borderline-SMOTE : A New Over-Sampling Method in Imbalanced Data Sets Learning, Advances in Intelligent Computing source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given

    Federated learning enables the collaborative training of a machine learning model and deep learning for mobile edge network optimization.

    “Federated learning enables the collaborative training of a machine learning model and deep learning for mobile edge network optimization.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Premisenone given

    For the prediction of the next word, cellular phone users may utilize a wide range of vocabulary.

    “For the prediction of the next word, cellular phone users may utilize a wide range of vocabulary.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationreasoning

    In such networks, communication requires more computations as compared to traditional data center environments.

    “In such networks, communication requires more computations as compared to traditional data center environments.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Interpretationnone given

    The amount of data on different edges may differ, and an underlying structure that reflects the interaction between devices and their associated distributions may exist.

    “Furthermore, the amount of data on different edges may differ, and an underlying structure that reflects the interaction between devices and their associated distributions may exist.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Benchmark LEAF provides some public federated datasets and evaluation framework.

    “A benchmark LEAF [66] provides some public federated datasets and evaluation framework, and other datasets and models used in existing publications in top tier conferences of the machine learning community throughout the past two years are summarized in Table 10 below.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [66] Caldas, S., Duddu, S.M.K., Wu, P., Li, T., Konečný, J., McMahan, H.B., Smith, V., and Talwalkar, A. (2018). LEAF: A Benchmark for Federated Settings. arXiv.

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    After careful screening and filtering, 105 papers were selected to adequately describe the research questions of this study.

    “After careful screening and filtering, 105 papers were selected to adequately describe the research questions of this study.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductnone given

    The search procedure was performed from April 2020 to May 2021, with the total initial number of papers being 7546 published in the duration of 2016 to 2020.

    “The search procedure was performed from April 2020 to May 2021, with the total initial number of papers being 7546 published in the duration of 2016 to 2020.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    This data generation paradigm challenges widely held i.i.d. principles in distributed optimization, raises the likelihood of stragglers, and may increase the complexity of analysis, modeling, and assessment.

    “This data generation paradigm challenges widely held i.i.d.principles in distributed optimization, raises the likelihood of stragglers, and may increase the complexity of analysis, modeling, and assessment.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    FL networks hypothetically encompassed a gigantic quantity of devices such as millions of notebooks and hand-held devices.

    “FL networks hypothetically encompassed a gigantic quantity of devices [17] (such as millions of notebooks and hand-held devices) and network communication may be computationally expensive and slower due to orders of magnitude.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [17] Li, 2020, Federated Learning: Challenges, Methods, and Future Directions, IEEE Signal Proces. Mag.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning

    The distribution of data from edge devices is based on their many uses.

    “As a result, the distribution of data from those edge devices is based on their many uses.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationnone given

    FL allows multiple nodes to form a joint learning model to address critical problems such as access rights, access to heterogeneous data, privacy, and security.

    “FL allows multiple nodes to form a joint learning model to address critical problems such as access rights, access to heterogeneous data, privacy, security, etc.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Federated EMNIST is a federated partition version that covers natural heterogeneity stemming from writing style.

    “Federated EMNIST 671,585 and 77,483 3400 EMNIST [66] is a federated partition version that covers natural heterogeneity stemming from writing style.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [66] Caldas, S., Duddu, S.M.K., Wu, P., Li, T., Konečný, J., McMahan, H.B., Smith, V., and Talwalkar, A. (2018). LEAF: A Benchmark for Federated Settings. arXiv.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisenone given

    Each edge node in FL trains a shared model using its local data.

    “Using their local data, each edge node in FL trains a shared model.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    ATV scheme under lower reputation thresholds provides higher accuracy than MSL.

    “ATV scheme under lower reputation thresholds provides higher accuracy than MSL.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [6] Kang, 2020, Reliable Federated Learning for Mobile Networks, IEEE Wirel. Commun. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    FedDT-TTS greatly improved transformer models' performance in federated learning, reducing total training time by 40%.

    “Their model greatly improved transformer models' performance in the federated learning, reducing total training time by 40%.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [96] Hong, Z., Wang, J., Qu, X., Liu, J., Zhao, C., and Xiao, J. (2021). Federated Learning with Dynamic Transformer for Text to Speech. arXiv. source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    In experiments, federated learning achieved 89% accuracy, while centralized DNN achieved 93% accuracy.

    “In the experiments, FL achieved 89% accuracy, while in centralized training 93% accuracy is for DNN.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [41] Mun, H., and Lee, Y. (2020). Internet Traffic Classification with Federated Learning. Electronics, 10. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    The method provides approximately 60% reduced VUEs large queue lengths, without additional power consumption, compared to an average queue-based baseline.

    “The method provides approximately 60% reduced VUEs large queue lengths, without additional power consumption, compared to an average queue-based baseline.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [9] Samarakoon, S., Bennis, M., Saad, W., and Debbah, M. (2018, January 9–13). Federated Learning for Ultra-Reliable Low-Latency V2V Communications. Proceedings of the 2018 IEEE Global Communications Conference (GLOBECOM), Abu Dhabi, United Arab Emirates. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    BFL is efficient in communication of autonomous vehicles.

    “BFL is efficient in communication of autonomous vehicles.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [114] Pokhrel, 2020, Federated Learning With Blockchain for Autonomous Vehicles: Analysis and Design Challenges, IEEE Trans. Commun. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    A non-IID image dataset containing 900+ images taken by 26 street cameras with seven types of objects is presented.

    “A non-IID image dataset containing 900+ images taken by 26 street cameras with seven types of objects is presented.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [61] Luo, J., Wu, X., Luo, Y., Huang, A., Huang, Y., Liu, Y., and Yang, Q. (2019). Real-World Image Datasets for Federated Learning. arXiv.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning

    FL has emerged as an innovative learning platform by enabling edge devices to train the model with their own data.

    “To mitigate the data privacy concerns, along with providing a transfer learning paradigm, FL has emerged as an innovative learning platform by enabling edge devices to train the model with their own data.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    DFNAS resulted in lesser computation and communication bandwidth with 92% test accuracy.

    “It resulted in lesser consume computation and communication bandwidth with 92% test accuracy.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [135] Garg, A., Saha, A.K., and Dutta, D. (2020). Direct Federated Neural Architecture Search. arXiv.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationnone given

    The domains of recommender systems, FinTech, and the autonomous industry can adapt FL, but the challenges these domains can encounter are statistical heterogeneity, system heterogeneity, data imbalance, resource allocation, and privacy.

    “The domains of recommender systems, FinTech, and the autonomous industry can adapt the FL, but the challenges these domains can encounter are statistical heterogeneity, system heterogeneity, data imbalance, resource allocation, and privacy.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationnone given

    The convergence of FedAvg is application-dependent.

    “For example, the convergence of its baseline aggregation algorithm, FedAvg, is application-dependent, and more refined methods for aggregation are worth exploring.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    FLIC can classify new applications on the fly with an accuracy of 88% under non-IID traffic across clients.

    “Tensorflow FLIC can classify new applications on the fly with an accuracy of 88% under non-IID traffic across clients.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [41] Mun, H., and Lee, Y. (2020). Internet Traffic Classification with Federated Learning. Electronics, 10. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    FASTGNN can provide similar performance compared with three baseline algorithms, where MAPE of FASTGNN is only 0.31% more than MAPE of STGCN.

    “FASTGNN can provide similar performance compared with the three baseline algorithms, where the MAPE of FASTGNN is only 0.31% more than MAPE of STGCN.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [37] Zhang, 2021, FASTGNN: A Topological Information Protected Federated Learning Approach for Traffic Speed Forecasting, IEEE Trans. Ind. Inform. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    MSL scheme performs the same as ATV scheme when the reputation is higher than 0.35.

    “MSL scheme performs the same as ATV scheme when the reputation is higher than 0.35.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [6] Kang, 2020, Reliable Federated Learning for Mobile Networks, IEEE Wirel. Commun. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    FedProx local optimizer allows better accuracy using DNN implementations.

    “FedProx local optimizer allows better accuracy using DNN implementations.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [146] Donevski, 2021, On Addressing Heterogeneity in Federated Learning for Autonomous Vehicles Connected to a Drone Orchestrator, Front. Commun. Netw.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationnone given

    The growing storage and computation capacity of edge nodes and fast communication such as 5G has revolutionized the way of machine learning in the modern era.

    “The growing storage and computation capacity of edge nodes, such as autonomous vehicles, smartphones, tablets, and fast communication such as 5G, FL has revolutionized the way of machine learning in the modern era.”
    Quote in the paperNot a substantial claim
94 of 94

  • The paper relies on this without support.

    The same FL approaches used on smartphones can be used for PCs, IoT, and other edge devices such as autonomous vehicles.

    “however, the same approaches can be used for PCs, IoT, and other edge devices such as autonomous vehicles.”
1 of 1

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Publication-year counts exceed the 105 selected studies

    Major

    The stated counts for 2016 through the first half of 2020 total 125, not 105. This makes the reported year-by-year breakdown inconsistent with the number of studies reviewed.

    Reported 105 · recomputed 125

  • CIFAR-100 client allocation accounts for only 50,000 records

    Minor

    The table reports 60,000 data items, but 500 clients with 100 records each account for 50,000.

    Reported 60000 · recomputed 50000

  • Challenges are directed to the algorithms table

    Minor

    The text says the challenges are summarized in Table 9, while Table 9 is labeled as algorithms and Table 11 is labeled as challenges.

3 of 3

Cites

Flagged first, then by citations

Research works printed but not linked: 37

  1. Reinsel, 2018, Data Age 2025: The Digitization of the World From Edge to Core, Int. Data Corp., 16, 28
  2. Ji, S., Saravirta, T., Pan, S., Long, G., and Walid, A. (2021). Emerging Trends in Federated Learning: From Model Fusion to Federated X Learning. arXiv.
  3. Zhao, Y., Zhao, J., Jiang, L., Tan, R., and Niyato, D. (2020). Mobile Edge Computing, Blockchain and Reputation-based Crowdsourcing IoT Federated Learning: A Secure, Decentralized and Privacy-preserving System. arXiv.
  4. Li, 2020, Federated Learning: Challenges, Methods, and Future Directions, IEEE Signal Proces. Mag., 37, 50
  5. Soto, J.C., Kyt, W., Jahn, M., Pullmann, J., Bonino, D., Pastrone, C., and Spirito, M. (December, January 30). Towards a Federation of Smart City Services. Proceedings of the International Conference on Recent Advances in Computer Systems (RACS 2015), Hail, Saudi Arabia.
  6. Bonawitz, 2019, Towards Federated Learning at Scale: System Design, Proc. Mach. Learn. Syst., 1, 374
  7. Tan, A.Z., Yu, H., Cui, L., and Yang, Q. (2021). Towards Personalized Federated Learning. arXiv.
  8. Li, 2020, Federated Optimization in Heterogeneous Networks, Proc. Mach. Learn. Syst., 2, 429
  9. Liu, M., Ho, S., Wang, M., Gao, L., Jin, Y., and Zhang, H. (2021). Federated Learning Meets Natural Language Processing: A Survey. arXiv.
  10. (2021, November 28). Federated Learning for Internet of Things and Big Data|Hindawi. Available online: https://www.hindawi.com/journals/wcmc/si/891247/.
  11. Elbir, A.M., Soner, B., and Coleri, S. (2020). Federated Learning for Vehicular Networks. arXiv.
  12. Bonawitz, K., Ivanov, V., Kreuter, B., Marcedone, A., McMahan, H.B., Patel, S., Ramage, D., Segal, A., and Seth, K. (2016). Practical Secure Aggregation for Federated Learning on User-Held Data. arXiv.
  13. Yang, C.S., So, J., He, C., Li, S., Yu, Q., and Avestimehr, S. (2021). LightSecAgg: Rethinking Secure Aggregation in Federated Learning. arXiv.
  14. He, C., Li, S., So, J., Zeng, X., Zhang, M., Wang, H., Shen, L., Yang, Y., Yang, Q., and Avestimehr, S. (2020). FedML: A Research Library and Benchmark for Federated Machine Learning. arXiv.
  15. Huang, B., Li, X., Song, Z., and Yang, X. (2021). FL-NTK: A Neural Tangent Kernel-based Framework for Federated Learning Convergence Analysis. arXiv.
  16. Garg, A., Saha, A.K., and Dutta, D. (2020). Direct Federated Neural Architecture Search. arXiv.
  17. Lee, J.W., Oh, J., Lim, S., Yun, S.Y., and Lee, J.G. (2020). TornadoAggregate: Accurate and Scalable Federated Learning via the Ring-Based Architecture. arXiv.
  18. Cheng, G., Chadha, K., and Duchi, J. (2021). Fine-tuning in Federated Learning: A simple but tough-to-beat baseline. arXiv.
  19. Sahu, A.K., Li, T., Sanjabi, M., Zaheer, M., Talwalkar, A., and Smith, V. (2018). On the convergence of federated optimization in heterogeneous networks. arXiv.
  20. Konečný, J., McMahan, H.B., Ramage, D., and Richtárik, P. (2016). Federated Optimization: Distributed Machine Learning for On-Device Intelligence. arXiv.
  21. (2022, January 19). FedSVRG Based Communication Efficient Scheme for Federated Learning in MEC Networks | Request PDF. Available online: https://www.researchgate.net/publication/352418092_FedSVRG_Based_Communication_Efficient_Scheme_for_Federated_Learning_in_MEC_Networks.
  22. Liu, Y., Kang, Y., Zhang, X., Li, L., Cheng, Y., Chen, T., Hong, M., and Yang, Q. (2019). A Communication Efficient Collaborative Learning Framework for Distributed Features. arXiv.
  23. Mohri, M., Sivek, G., and Suresh, A.T. (2019, January 9–15). Agnostic Federated Learning. Proceedings of the 36th International Conference on Machine Learning, Long Beach, CA, USA.
  24. Afonin, A., and Karimireddy, S.P. (2021). Towards Model Agnostic Federated Learning Using Knowledge Distillation. arXiv.
  25. Wang, H., Yurochkin, M., Sun, Y., Papailiopoulos, D., and Khazaeni, Y. (2020). Federated Learning with Matched Averaging. arXiv.
  26. Liu, Y., Liu, Y., Liu, Z., Liang, Y., Meng, C., Zhang, J., and Zheng, Y. (2020). Federated Forest. IEEE Trans. Big Data.
  27. Dong, T., Li, S., Qiu, H., and Lu, J. (2022). An Interpretable Federated Learning-based Network Intrusion Detection Framework. arXiv.
  28. (2022, January 25). An Example of Implementing FL for Linear Regression.|Download Scientific Diagram. Available online: https://www.researchgate.net/figure/An-example-of-implementing-FL-for-linear-regression_fig3_346038614.
  29. Hardy, S., Henecka, W., Ivey-Law, H., Nock, R., Patrini, G., Smith, G., and Thorne, B. (2017). Private federated learning on vertically partitioned data via entity resolution and additively homomorphic encryption. arXiv.
  30. Chen, F., Luo, M., Dong, Z., Li, Z., and He, X. (2018). Federated Meta-Learning with Fast Convergence and Efficient Communication. arXiv.
  31. Zhou, F., Wu, B., and Li, Z. (2018). Deep Meta-Learning: Learning to Learn in the Concept Space. arXiv.
  32. (2022, January 25). Federated Meta-Learning for Recommendation—arXiv Vanity. Available online: https://www.arxiv-vanity.com/papers/1802.07876/.
  33. Pye, S.K., and Yu, H. (2021). Personalized Federated Learning : A Combinational Approach. arXiv.
  34. Liang, X., Liu, Y., Chen, T., Liu, M., and Yang, Q. (2019). Federated Transfer Reinforcement Learning for Autonomous Driving. arXiv.
  35. Basnayake, V. (2019). Federated Learning for Enhanced Sensor Realiabity of Automated Wireless Networks. [Master’s Thesis, University of Oulu].
  36. Ren, J., Ni, W., Nie, G., and Tian, H. (2021). Research on Resource Allocation for Efficient Federated Learning. arXiv.
  37. Zheng, Z., Zhou, Y., Sun, Y., Wang, Z., Liu, B., and Li, K. (2021). Federated Learning in Smart Cities: A Comprehensive Survey. arXiv.

Cited by

Sorted by citations

Related

Authors
Shaheen, M., Farooq, M. S., Umer, T., & Kim, B.-S.
Affiliations
  • COMSATS University Islamabad (PK)
  • University of Management and Technology (PK)
  • Hongik University (KR)
Corresponding author
Kim, Byung-Seo
ORCID
Venue
Electronics
Publisher
MDPI AG
Published
2022-02-21
Volume
11
Issue
4
Pages
670
ISSN
2079-9292 (ISSN-L)
DOI
10.3390/electronics11040670
OpenAlex
W4213010368
Cited by
157
References
151
Citations by year
2026: 25 · 2025: 34 · 2024: 45 · 2023: 42 · 2022: 11
FWCI
10.86
Type
journal article
Language
en
Open access
Licence
Locations
  • Electronics (publishedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
National Research Foundation of Korea (BK21 FOUR (F21YY8102068))
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data
  • Physical Sciences › Computer Science › Computer Networks and Communications › IoT and Edge/Fog Computing
  • Physical Sciences › Computer Science › Artificial Intelligence › Domain Adaptation and Few-Shot Learning
Keywords
federated learning, edge networks, edge computing, Internet of Things, medical applications, autonomous vehicles, distributed datasets, statistical heterogeneity, system heterogeneity, class imbalance, resource allocation, privacy

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-09.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-09.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review