2020
DOI: 10.1007/978-3-030-63076-8_10
|View full text |Cite
|
Sign up to set email alerts
|

Efficient and Fair Data Valuation for Horizontal Federated Learning

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
31
0

Year Published

2021
2021
2024
2024

Publication Types

Select...
4
2
1

Relationship

0
7

Authors

Journals

citations
Cited by 37 publications
(34 citation statements)
references
References 3 publications
0
31
0
Order By: Relevance
“…The diference is that MR calculates SVs during every round of FL model training, while OR calculates SV only once after all the training rounds are complete. In [26], the authors proposed the Truncated MR (TMR) approach, which extends MR by eliminating the entire rounds with decay factor λ values lower than a pre-deined threshold. These three schemes signiicantly improve the eiciency of estimating SVs.…”
Section: Fl Participant Contribution Evaluation Via Gradient Shapley ...mentioning
confidence: 99%
See 3 more Smart Citations
“…The diference is that MR calculates SVs during every round of FL model training, while OR calculates SV only once after all the training rounds are complete. In [26], the authors proposed the Truncated MR (TMR) approach, which extends MR by eliminating the entire rounds with decay factor λ values lower than a pre-deined threshold. These three schemes signiicantly improve the eiciency of estimating SVs.…”
Section: Fl Participant Contribution Evaluation Via Gradient Shapley ...mentioning
confidence: 99%
“…We analyze the MR approach [26], which combines the canonical SV calculation and gradient-based model reconstruction. We adopt the non-i.i.d FL data silos setting as described in the experiment section with 10 participants and MNIST dataset [11].…”
Section: Preliminariesmentioning
confidence: 99%
See 2 more Smart Citations
“…It is essential to understand how to evaluate the contributions of data owners based on federated learning. Wei et al, 2020;Wang et al, 2020).…”
Section: Introductionmentioning
confidence: 99%