1 of 94

Client Heterogeneity in Federated Learning Systems

Angelo RODIO

July 3, 2024

Centre Inria d'Université Côte d'Azur

2 of 94

Outline

  • Introduction: Motivations, Challenges & Contributions
  • A focus on Client Participation Heterogeneity
    • Part 1. Heterogeneous and Correlated Client Participation
    • Part 2. Leveraging Stale Updates for Non-Participating Clients
  • Part 3. Other Contributions
  • Conclusion & Future work

3 of 94

Motivations: "ML models need data"

WE ARE HERE

Hestness, et al. "Deep learning scaling is predictable, empirically.", arXiv, 2017

error: smaller is better

dataset size: note log axis

Villalobos, et al. "Position: Will we run out of data? Limits of LLM scaling based on human-generated data." ICML, 2024

1

4 of 94

Motivations: "Data production at end-devices"

ML model

Data

Clients

Cloud server

2

5 of 94

A shift of paradigm:

Data

Clients

ML model

Data

Centralized optimization

loss function

data sample

from centralized data

to decentralized data

Cloud server

Data may be sensitive

Sending data may be costly

3

6 of 94

Federated Learning as a solution

[McM’16; Kon’16]

Data

Clients

where

loss function

data sample

global objective

local objective

local objective

ML model

Cloud server

4

7 of 94

Federated Learning as a solution

5

Cross-silo FL

Cross-device FL

Hospitals

Organizations

smartphones, voice assistants, social media

Github: Pold87/academic-keyword-occurrence

[McM’16; Kon’16]

8 of 94

A baseline algorithm: FedAvg

Data

Clients

Cloud server

initial model

Objective:

6

[McM’16; Kon’16; Red’19]

9 of 94

A baseline algorithm: FedAvg

Clients

Cloud server

client optimization

Objective:

6

[McM’16; Kon’16; Red’19]

10 of 94

A baseline algorithm: FedAvg

Clients

Cloud server

client update

Objective:

6

[McM’16; Kon’16; Red’19]

11 of 94

A baseline algorithm: FedAvg

Clients

Cloud server

server optimization

Objective:

6

server aggregation

set of participating clients

client update

[McM’16; Kon’16; Red’19]

12 of 94

Statistical heterogeneity

System heterogeneity

Heterogeneous network

Heterogeneous hardware

Heterogeneous power, etc

Challenges & Contributions

7

+

from

from

from

Client participation heterogeneity

results into

[INFOCOM’23]

[TON’23]

[SIGMETRICS SRC’24]

[arXiv’24]

[WPMC’23]

[arXiv’24]

*bold: first author

Other challenges:

- personalization�- fairness�- robustness�- privacy

PART 1

PART 2

PART 3 (overview)

[Wang’20; Kai’20; Li’20]

13 of 94

In this presentation

8

Statistical heterogeneity

System heterogeneity

Heterogeneous network

Heterogeneous hardware

Heterogeneous power, etc

+

from

from

from

Client participation heterogeneity

results into

14 of 94

Client participation heterogeneity

Publications

In this presentation

Problem

Key idea

Algorithm

variance from correlation

variance from low participation

leverage optim-bias tradeoff

leverage stale client updates

Correlation-Aware FL (CaFed)

Staleness-Aware FL (FedStale)

8

[IEEE INFOCOM 2023]

[IEEE/ACM Trans. On Net. 2023]

[ACM SIGMETRICS SRC 2024]

[European Conf. AI 2024]

15 of 94

PART 1

Heterogeneous and correlated client participation

  1. Rodio, Faticanti, Marfoq, Neglia, and Leonardi. "Federated Learning under Heterogeneous and Correlated Client Availability." In IEEE INFOCOM, 2023

  • Rodio, Faticanti, Marfoq, Neglia, and Leonardi. "Federated Learning under Heterogeneous and Correlated Client Availability." In IEEE/ACM Trans. On Net., 2023

8

16 of 94

Heterogeneous and correlated client participation

Common Assumption [Li’19, Kar’20]

Communication rounds

Clients

Communication rounds

Our work

client participation

sampled client

In our work, participation is heterogeneous, correlated over time (temporal) &�among clients (spatial)

temporal correlation

spatial correlation

[Wan’20, Red’21]

temporal correlation

spatial correlation

When the population �is large, the server can �work with K/N clients.

Common assumption:

9

17 of 94

Assumption to model the heterogeneous �and correlated client participation

  • : set of participating clients at time t
  • is a discrete-time Markov chain
    • transition matrix
    • stationary distribution
    • largest 2nd eigenvalue

E.g., client participations are independent:

avg. participation

correlation

Client i

not participating

participating

10

18 of 94

Related work

Spatial correlation

Temporal correlation

  • Cyclic participation [Eic’19; Zhu’20; Cho’23]:�deterministic participation patterns
  • F3AST [Rib’23]: �uncorrelated/positively correlated
  • Markov-chain GD [Sun’18; Doa’20]:�correlation in data sampling

  • This work:�correlation in client participation

11

19 of 94

General FL algorithm

server aggregation

server optimization

Objective:

WE CAN CHOOSE

Clients

Cloud server

12

20 of 94

Problem: client heterogeneity

12

Before, we considered IID data distributions…

"happy" & "sad"

"happy" & "sad"

21 of 94

Problem: client heterogeneity

Statistical heterogeneity

"sad" client

"happy" client

12

22 of 94

Problem: client heterogeneity

Statistical heterogeneity

Participation heterogeneity

Model Bias�in favor of the more participating client

+

=

"sad" client

"happy" client

12

23 of 94

Problem: client heterogeneity

13

24 of 94

Problem: client heterogeneity

13

25 of 94

Problem: client heterogeneity

biased objective

14

26 of 94

Bias error (def.)

Bias Error

where we should have converged

where we eventually converge

15

27 of 94

Bias error (theorem)

total variation

statistical heterogeneity

15

28 of 94

Bias error (guideline)

Unbiased FedAvg

total variation

statistical heterogeneity

Guideline 1:

15

29 of 94

Unbiased FedAvg

16

30 of 94

Unbiased FedAvg

16

31 of 94

  • Q: Is Unbiased FedAvg enough?

Unbiased FedAvg

  • A: An Optimization-Bias tradeoff
    • slower convergence
    • faster convergence

16

  • Q: What about correlation?

32 of 94

  • Q: Is Unbiased FedAvg enough?

Unbiased FedAvg

  • A: An Optimization-Bias tradeoff
    • slower convergence
    • faster convergence

16

  • Q: What about correlation?

33 of 94

Optimization error

No Correlation ( )

17

Client 1 ONLY

Client 2 PARTICIPATES CONSECUTIVELY

Large variability

Low variability

Correlation ( )

34 of 94

Optimization error (def.)

where we actually stop

where we eventually converge

18

: #rounds

35 of 94

: #rounds

avg. participation

correlation

Optimization error (theorem)

WE CAN CHOOSE

WE CAN CONTROL

first convergence analysis!

18

36 of 94

Exclude ( ) "more correlated" clients

correlation

Optimization error (guidelines)

WE CAN CHOOSE

WE CAN CONTROL

Exclude ( ) "less participating" clients

Guideline 2:

Guideline 3:

18

: #rounds

avg. participation

37 of 94

Minimize the green term for large

: #rounds

avg. availability

correlation

Optimization error (guidelines)

Minimize the blue term large for large

WE CAN CHOOSE

WE CAN CONTROL

38 of 94

Total error (def.)

Total Error

where we actually stop

where we should have converged

39 of 94

Total error (theorem)

Total Error

40 of 94

Correlation-Aware FL (CaFed)

1) The server

  • Estimates

19

41 of 94

Correlation-Aware FL (CaFed)

  • Initializes

1) The server

  • Estimates

(G1)

19

42 of 94

Correlation-Aware FL (CaFed)

2) For , starting from the

"Less Participating", "More Correlated" clients:

(G2)

(G3)

19

43 of 94

  • Try to exclude client ( )

Correlation-Aware FL (CaFed)

2) For , starting from the

"Less Participating", "More Correlated" clients:

(G2)

(G3)

19

44 of 94

Correlation-Aware FL (CaFed)

  • If decreases:

confirm decision

  • Try to exclude client ( )

2) For , starting from the

"Less Participating", "More Correlated" clients:

(G2)

(G3)

19

45 of 94

  • Population of N=100 clients, divided in

Experimental setting

  • Optimization-Bias tradeoff

More Participating

Minimize the optimization error

Faster convergence to a biased objective

 Unbiased

Minimize the bias error

Slower convergence to target objective

20

46 of 94

  • CaFed w/ temporal and spatial correlation

Experimental results

CaFed entirely excludes "Less Participating", "More Correlated" clients

21

When data distribution is intra-group homogeneous, inter-group homogeneous

CASE 1

47 of 94

  • CaFed w/ temporal and spatial correlation

Experimental results

CaFed excludes some of the "Less Participating", "More Correlated" clients

23

When data distribution is intra-group homogeneous, inter-group heterogeneous

CASE 2

48 of 94

  • CaFed w/ temporal and spatial correlation

Experimental results

21

spatial correlation

temporal correlation

49 of 94

  • CaFed w/ temporal and spatial correlation

Experimental results

CaFed excludes some "Less Participating", "More Correlated" clients adaptively over time

21

50 of 94

  • CaFed w/ temporal and spatial correlation

CaFed achieves higher accuracy within training time

Experimental results

21

CIFAR-10

51 of 94

  • Client participation is correlated over time and space
  • First convergence analysis under Markovian client participation
  • Correlation-Aware (CaFed) algorithm adaptively excludes

Part 1. Conclusions

the "less participating", "more correlated" clients

22

52 of 94

PART 2

Leveraging Stale Updates for Non-Participating Clients

  1. Rodio. "The multiple facets of Variance Reduction in Federated Learning." In ACM SIGMETRICS Student Research Competition (SRC), 2024�
  2. Rodio and Neglia. "FedStale: leveraging Stale Clients Updates in Federated Learning." arXiv preprint, 2024

23

53 of 94

Variance from participation heterogeneity

Unbiased FedAvg [Rod’23; Wang’22,24]

set of �participating �clients

(asymptotic)�participation probability of client i

Cloud server

24

54 of 94

Cloud server

set of �participating �clients

Variance from participation heterogeneity

24

(asymptotic)�participation probability of client i

Unbiased FedAvg [Rod’23; Wang’22,24]

55 of 94

Variance from participation heterogeneity

Client 1 ONLY

Client 2 PARTICIPATES

Client 1 ONLY

Client 2 PARTICIPATES

25

Large variance

U-FedAvg

Q: Why?

56 of 94

Q: Can we reduce this variance?

U-FedAvg (Convergence)

Upper bound. Under Assumptions 1-3, for smooth, non-convex objectives:

26

57 of 94

Variance Reduction. Leverage stale updates �for non-participating clients [Gu’21; Jhu’22]

Cloud server

memory term: stale update of client i

27

[Yang’22; Yan’24]

Related work

58 of 94

27

memory term: stale update of client i

Variance Reduction. Leverage stale updates �for non-participating clients [Gu’21; Jhu’22]

[Yang’22; Yan’24]

Related work

Cloud server

59 of 94

Q: Why?

Client 2 quits participation

Suboptimal trajectory

Variance Reduction

Client 1 ONLY

IF Client 2 HAD PARTICIPATED

Ideal trajectory

Stale Update

Ideal trajectory

28

U-FedVARP

U-FedAvg

60 of 94

Convergence

Upper bound. Under Assumptions 1-3, for smooth, non-convex objectives:

Lower bound. The term is necessary:

Our findings.

29

61 of 94

FedStale

30

62 of 94

30

Staleness-Aware FL (FedStale)

63 of 94

Q: Why does it work?

Staleness-Aware FL (FedStale)

A convex combination of �"fresh" and "stale" updates

Cloud server

31

weight to stale updates

64 of 94

  •  

Convergence

Upper bound. Under Assumptions 1-3, for smooth, non-convex objectives:

Minimizing the bound:

32

65 of 94

Experimental results

33

66 of 94

  • Participation heterogeneity introduces model variability
  • Variance Reduction must aggregate updates of varying staleness
  • Staleness-Aware (FedStale) algorithm combines "fresh" � and "stale" updates, with theoretical guarantees

Part 2. Conclusions

34

67 of 94

Client participation heterogeneity

Publications

Comparison

Problem

Key idea

Algorithm

variance from correlation

variance from participation

leverage optim-bias tradeoff

leverage stale client updates

Correlation-Aware FL (CaFed)

Staleness-Aware FL (FedStale)

[IEEE INFOCOM 2023]

[IEEE/ACM Trans. On Net. 2023]

[ACM SIGMETRICS SRC 2024]

[arXiv 2024]

PART 1

PART 2

35

68 of 94

PART 3

  • Other contributions
  1. Rodio, Neglia, Busacca, Mangione, Palazzo, Restuccia, and Tinnirello. "Federated Learning with Packet Losses." In WPMC, 2023�
  2. Kaplan, Si Salem, Rodio, Xu, and Neglia. "Federated Learning for Cooperative Inference Systems: The Case of Early Exit Networks." arXiv preprint, 2024

36

69 of 94

Statistical heterogeneity

System heterogeneity

Heterogeneous network

Heterogeneous hardware

Heterogeneous power, etc

Challenges & Contributions

+

from

from

from

Client participation heterogeneity

results into

[INFOCOM’23]

[TON’23]

[SIGMETRICS SRC’24]

[ArXiv’24]

[WPMC’23]

[arXiv’24]

PART 3 (overview)

37

70 of 94

Heterogeneous network resources

39

80 rounds

+10%

residual errors

  • Diverse communication channels degrade training performance
  • Packet loss-aware �(UPGA-PL) algorithm
  • Robustness of FL algorithms to intermediate error (packet loss)

71 of 94

Heterogeneous network resources

38

  • Diverse communication channels degrade training performance
  • Packet loss-aware algorithm,�with theoretical guarantees
  • Robustness of FL algorithms to �errors (e.g., packet losses)

[YAN’20; ERI’21]

[CHE’21; YE’22]

[XIN’16; MIS’19]

[MAL’16; PHI’19]

[WPMC’23]

72 of 94

Heterogeneous hardware resources

39

  • Clients: end-devices, �edge servers, cloud
  • Jointly train models �of different sizes
  • FL Algorithm, with theoretical guarantees

[REN’23; SIS’24]

[ArXiv’24]

73 of 94

74 of 94

Client Participation in Federated Learning Systems

Incentivizing client participation through personalization[MAN’20; DEN’20]

Decentralized or �semi-decentralized topologies[KOL’20; KIN’21]

Future Research Directions

40

Client

Federation

What do �I gain?

Personalization:

Server

Client

Neighbors

I can’t talk to the server

but I can talk to my neighbors

Remote�server

75 of 94

Heterogeneous network resouces

Heterogeneous energy resouces

Heterogeneous client participation

Incentivizing client participation through personalization

Decentralized or �semi-decentralized topologies

Retransmissions? When? How many?

Correlated client participation driven by energy contraints

Future Research Directions

35

76 of 94

References

41

Machine Learning models are data-hungry��[HES’17] J. Hestness et al., “Deep Learning Scaling is Predictable, Empirically.” arXiv, Dec. 01, 2017.

[VIL’24] P. Villalobos, A. Ho, J. Sevilla, T. Besiroglu, L. Heim, and M. Hobbhahn, “Will we run out of data? Limits of LLM scaling based on human-generated data,” presented at the Forty-first International Conference on Machine Learning, Jun. 2024��Federated Learning

[MCM’17] B. McMahan, E. Moore, D. Ramage, S. Hampson, and B. A. y Arcas, “Communication-Efficient Learning of Deep Networks from Decentralized Data,” in Proceedings of the 20th International Conference on Artificial Intelligence and Statistics, PMLR, Apr. 2017, pp. 1273–1282.

[KON’17] J. Konečný, H. B. McMahan, F. X. Yu, P. Richtárik, A. T. Suresh, and D. Bacon, “Federated Learning: Strategies for Improving Communication Efficiency.” arXiv, Oct. 2017.

[RED’21] S. J. Reddi et al., “Adaptive Federated Optimization,” in International Conference on Learning Representations, 2021.�

77 of 94

Federated Learning Surveys

[WAN’21] J. Wang et al., “A Field Guide to Federated Optimization.” arXiv, Jul. 14, 2021. doi: 10.48550/arXiv.2107.06917.

[KAI’21] P. Kairouz et al., “Advances and Open Problems in Federated Learning,” Foundations and Trends® in Machine Learning, vol. 14, no. 1–2, pp. 1–210, Jun. 2021, doi: 10.1561/2200000083.

[LI’20] T. Li, A. K. Sahu, A. Talwalkar, and V. Smith, “Federated Learning: Challenges, Methods, and Future Directions,” IEEE Signal Processing Magazine, vol. 37, no. 3, pp. 50–60, May 2020, doi: 10.1109/MSP.2020.2975749.

�Convergence analysis of FL algorithms��[LI’20] X. Li, K. Huang, W. Yang, S. Wang, and Z. Zhang, “On the Convergence of FedAvg on Non-IID Data,” in International Conference on Learning Representations, Apr. 2023.

[WAN’20] J. Wang, Q. Liu, H. Liang, G. Joshi, and H. V. Poor, “Tackling the Objective Inconsistency Problem in Heterogeneous Federated Optimization,” in Advances in Neural Information Processing Systems, Curran Associates, Inc., 2020.

[KAR’20] S. P. Karimireddy, S. Kale, M. Mohri, S. Reddi, S. Stich, and A. T. Suresh, “SCAFFOLD: Stochastic Controlled Averaging for Federated Learning,” in Proceedings of the 37th International Conference on Machine Learning, PMLR, Nov. 2020.

42

78 of 94

Federated Learning with Cyclic Client Participation

[EIC’19] H. Eichner, T. Koren, B. Mcmahan, N. Srebro, and K. Talwar, “Semi-Cyclic Stochastic Gradient Descent,” in Proceedings of the 36th International Conference on Machine Learning, PMLR, May 2019, pp. 1764–1773.

[ZHU’21] C. Zhu, Z. Xu, M. Chen, J. Konečnỳ, A. Hard, and T. Goldstein, “Diurnal or nocturnal? Federated learning from periodically shifting distributions,” in NeurIPS 2021 Workshop on Distribution Shifts, 2021.

[CHO’23] Y. J. Cho, P. Sharma, G. Joshi, Z. Xu, S. Kale, and T. Zhang, “On the Convergence of Federated Averaging with Cyclic Client Participation,” in Proceedings of the 40th International Conference on Machine Learning, PMLR, Jul. 2023.��Federated Learning with Temporal and Spatial Correlation

[RIB’20] M. Ribero, H. Vikalo, and G. de Veciana, “Federated Learning Under Intermittent Client Availability and Time-Varying Communication Constraints,” IEEE Journal of Selected Topics in Signal Processing, vol. 17, no. 1, pp. 98–111, Jan. 2023, doi: 10.1109/JSTSP.2022.3224590.

[SUN’18] T. Sun, Y. Sun, and W. Yin, “On Markov Chain Gradient Descent,” in Advances in Neural Information Processing Systems, Curran Associates, Inc., 2018.

[DOA’20] T. T. Doan, “Local Stochastic Approximation: A Unified View of Federated Learning and Distributed Multi-Task Reinforcement Learning Algorithms,” arXiv:2006.13460, Jun. 2020, doi: 10.48550/arXiv.2006.13460.��

43

79 of 94

Unbias FedAvg due to Heterogeneous Client Participation

[WAN’22] S. Wang and M. Ji, “A Unified Analysis of Federated Learning with Arbitrary Client Participation,” Advances in Neural Information Processing Systems, vol. 35, pp. 19124–19137, Dec. 2022.

[WAN’24] S. Wang and M. Ji, “A Lightweight Method for Tackling Unknown Participation Statistics in Federated Averaging.” arXiv, Jan. 2024.��Federated Learning with Variance Reduction

[GU’21] X. Gu, K. Huang, J. Zhang, and L. Huang, “Fast Federated Learning in the Presence of Arbitrary Device Unavailability,” in Advances in Neural Information Processing Systems, Curran Associates, Inc., 2021, pp. 12052–12064.

[JHU’22] D. Jhunjhunwala, P. Sharma, A. Nagarkatti, and G. Joshi, “Fedvarp: Tackling the Variance due to Partial Client Participation in Federated Learning,” in Proceedings of the Thirty-Eighth Conference on Uncertainty in Artificial Intelligence, PMLR, Aug. 2022, pp. 906–916.

[YAN’22] H. Yang, X. Zhang, P. Khanduri, and J. Liu, “Anarchic Federated Learning,” in Proceedings of the 39th International Conference on Machine Learning, PMLR, Jun. 2022, pp. 25331–25363.

[YAN’24] Y. Yan et al., “Federated Optimization Under Intermittent Client Availability,” INFORMS Journal on Computing, vol. 36, no. 1, pp. 185–202, Jan. 2024, doi: 10.1287/ijoc.2022.0057.

44

80 of 94

Federated Learning with Lossy Communication Channels

[ERI’21] M. C. Eriş, B. Kantarci, and S. Oktug, “Unveiling the Wireless Network Limitations in Federated Learning,” in 2021 IEEE International Symposium on Dynamic Spectrum Access Networks (DySPAN), Dec. 2021.

[YAN’20] H. H. Yang, Z. Liu, T. Q. S. Quek, and H. V. Poor, “Scheduling Policies for Federated Learning in Wireless Networks,” IEEE Transactions on Communications, vol. 68, no. 1, pp. 317–333, Jan. 2020.

[YE’22] H. Ye, L. Liang, and G. Y. Li, “Decentralized Federated Learning With Unreliable Communications,” IEEE Journal of Selected Topics in Signal Processing, vol. 16, no. 3, pp. 487–500, Apr. 2022.

[CHE’21] M. Chen, Z. Yang, W. Saad, C. Yin, H. V. Poor, and S. Cui, “A Joint Learning and Communications Framework for Federated Learning Over Wireless Networks,” IEEE Transactions on Wireless Communications, Jan. 2021.��Federated Learning is robust to intermediate errors�

[XIN’16] E. P. Xing, Q. Ho, P. Xie, and D. Wei, “Strategies and Principles of Distributed Machine Learning on Big Data,” Engineering, vol. 2, no. 2, pp. 179–195, Jun. 2016.��[MIS’23] K. Mishchenko, E. Gorbunov, M. Takáč, and P. Richtárik, “Distributed Learning with Compressed Gradient Differences.” arXiv, Dec. 28, 2023. doi: 10.48550/arXiv.1901.09269.

45

81 of 94

[MIS’22] K. Mishchenko, G. Malinovsky, S. Stich, and P. Richtarik, “ProxSkip: Yes! Local Gradient Steps Provably Lead to Communication Acceleration! Finally!,” in Proceedings of the 39th International Conference on Machine Learning, Jun. 2022.

[PHI’22] C. Philippenko and A. Dieuleveut, “Bidirectional compression in heterogeneous settings for distributed or federated learning with partial participation: tight convergence guarantees.” arXiv, Jun. 19, 2022. doi: 10.48550/arXiv.2006.14591.��Cooperative Inference Systems for Distributed Machine Learning

[REN’23] W.-Q. Ren et al., “A Survey on Collaborative DNN Inference for Edge Intelligence,” Mach. Intell. Res., vol. 20, no. 3, pp. 370–395, Jun. 2023, doi: 10.1007/s11633-022-1391-7.

[SIS’24] T. Si Salem, G. Castellano, G. Neglia, F. Pianese, and A. Araldo, “Toward Inference Delivery Networks: Distributing Machine Learning With Optimality Guarantees,” IEEE/ACM Transactions on Networking, vol. 32, no. 1, pp. 859–873, Feb. 2024.��Incentives to Federated Learning through Personalization

[DON’20] K. Donahue and J. Kleinberg, “Model-sharing Games: Analyzing Federated Learning Under Voluntary Participation.” arXiv, Dec. 17, 2020. doi: 10.48550/arXiv.2010.00753.

[GRI’21] F. Grimberg, M.-A. Hartley, S. P. Karimireddy, and M. Jaggi, “Optimal Model Averaging: Towards Personalized Collaborative Learning.” arXiv, Oct. 25, 2021. doi: 10.48550/arXiv.2110.12946.

����

46

82 of 94

[MAN’20] Y. Mansour, M. Mohri, J. Ro, and A. T. Suresh, “Three Approaches for Personalization with Applications to Federated Learning.” arXiv, Jul. 19, 2020. doi: 10.48550/arXiv.2002.10619.

�[DEN’20] Y. Deng, M. M. Kamani, and M. Mahdavi, “Adaptive Personalized Federated Learning.” arXiv, Nov. 05, 2020. doi: 10.48550/arXiv.2003.13461.��Decentralized and Semi-Decentralized Topologies for Federated Learning

[KOL’20] A. Koloskova, N. Loizou, S. Boreiri, M. Jaggi, and S. Stich, “A Unified Theory of Decentralized SGD with Changing Topology and Local Updates,” in Proceedings of the 37th International Conference on Machine Learning, PMLR, Nov. 2020.

[LIN’21] F. P.-C. Lin, S. Hosseinalipour, S. S. Azam, C. G. Brinton, and N. Michelusi, “Semi-Decentralized Federated Learning With Cooperative D2D Local Model Aggregations,” IEEE Journal on Selected Areas in Communications, vol. 39, no. 12, pp. 3851–3869, Dec. 2021, doi: 10.1109/JSAC.2021.3118344.

[COS’23] M. Costantini, G. Neglia, and T. Spyropoulos, “FedDec: Peer-to-peer Aided Federated Learning.” arXiv, Jun. 11, 2023. doi: 10.48550/arXiv.2306.06715.

[CHE’24] E. Chen, S. Wang, and C. G. Brinton, “Taming Subnet-Drift in D2D-Enabled Fog Learning: A Hierarchical Gradient Tracking Approach.” arXiv, Jan. 09, 2024. doi: 10.48550/arXiv.2312.04728.

47

83 of 94

Problem

Publications

system heterogeneity

variance from participation

Take-home

variance from correlation

leverage optim-bias tradeoff

leverage stale client updates

Correlation-Aware FL (CaFed)

Staleness-Aware FL (FedStale)

[IEEE INFOCOM 2023]

[IEEE/ACM Trans. On Net. 2023]

[ACM SIGMETRICS SRC 2024]

[arXiv 2024]

PART 1

PART 2

PART 3

Heterogeneity-Aware FL

convergence analysis

[WPMC 2023]

[arXiv 2024]

Key idea

Algorithm

48

84 of 94

BACKUP

85 of 94

Client participation heterogeneity

Approach

Comparison

Problem

Assumptions

Algorithm

variance from correlation

variance from participation

smoothness; strong-convexity

only smoothness

strongly relies on bound

loosely relies on bound

PART 1

PART 2

top-down

bottom-up

86 of 94

87 of 94

88 of 94

Characterization of 2-state MC

89 of 94

Proof intuition

90 of 94

Participation Estimation (CaFed)

91 of 94

Participation Estimation (FedStale)

92 of 94

More experiments (FedStale)

93 of 94

Future Research Directions (Personalization)

94 of 94