LIA: Privacy-Preserving Data Quality Evaluation in Federated Learning Using a Lazy Influence Approximation
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2022
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866912133131796480 |
|---|---|
| author | Rokvic, Ljubomir Danassis, Panayiotis Karimireddy, Sai Praneeth Faltings, Boi |
| author_facet | Rokvic, Ljubomir Danassis, Panayiotis Karimireddy, Sai Praneeth Faltings, Boi |
| contents | In Federated Learning, it is crucial to handle low-quality, corrupted, or malicious data. However, traditional data valuation methods are not suitable due to privacy concerns. To address this, we propose a simple yet effective approach that utilizes a new influence approximation called "lazy influence" to filter and score data while preserving privacy. To do this, each participant uses their own data to estimate the influence of another participant's batch and sends a differentially private obfuscated score to the central coordinator. Our method has been shown to successfully filter out biased and corrupted data in various simulated and real-world settings, achieving a recall rate of over $>90\%$ (sometimes up to $100\%$) while maintaining strong differential privacy guarantees with $\varepsilon \leq 1$. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2205_11518 |
| institution | arXiv |
| publishDate | 2022 |
| record_format | arxiv |
| spellingShingle | LIA: Privacy-Preserving Data Quality Evaluation in Federated Learning Using a Lazy Influence Approximation Rokvic, Ljubomir Danassis, Panayiotis Karimireddy, Sai Praneeth Faltings, Boi Cryptography and Security Artificial Intelligence Machine Learning In Federated Learning, it is crucial to handle low-quality, corrupted, or malicious data. However, traditional data valuation methods are not suitable due to privacy concerns. To address this, we propose a simple yet effective approach that utilizes a new influence approximation called "lazy influence" to filter and score data while preserving privacy. To do this, each participant uses their own data to estimate the influence of another participant's batch and sends a differentially private obfuscated score to the central coordinator. Our method has been shown to successfully filter out biased and corrupted data in various simulated and real-world settings, achieving a recall rate of over $>90\%$ (sometimes up to $100\%$) while maintaining strong differential privacy guarantees with $\varepsilon \leq 1$. |
| title | LIA: Privacy-Preserving Data Quality Evaluation in Federated Learning Using a Lazy Influence Approximation |
| topic | Cryptography and Security Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2205.11518 |