Unveiling Privacy, Memorization, and Input Curvature Links
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866909123209068544 |
|---|---|
| author | Ravikumar, Deepak Soufleri, Efstathia Hashemi, Abolfazl Roy, Kaushik |
| author_facet | Ravikumar, Deepak Soufleri, Efstathia Hashemi, Abolfazl Roy, Kaushik |
| contents | Deep Neural Nets (DNNs) have become a pervasive tool for solving many emerging problems. However, they tend to overfit to and memorize the training set. Memorization is of keen interest since it is closely related to several concepts such as generalization, noisy learning, and privacy. To study memorization, Feldman (2019) proposed a formal score, however its computational requirements limit its practical use. Recent research has shown empirical evidence linking input loss curvature (measured by the trace of the loss Hessian w.r.t inputs) and memorization. It was shown to be ~3 orders of magnitude more efficient than calculating the memorization score. However, there is a lack of theoretical understanding linking memorization with input loss curvature. In this paper, we not only investigate this connection but also extend our analysis to establish theoretical links between differential privacy, memorization, and input loss curvature. First, we derive an upper bound on memorization characterized by both differential privacy and input loss curvature. Second, we present a novel insight showing that input loss curvature is upper-bounded by the differential privacy parameter. Our theoretical findings are further empirically validated using deep models on CIFAR and ImageNet datasets, showing a strong correlation between our theoretical predictions and results observed in practice. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_18726 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Unveiling Privacy, Memorization, and Input Curvature Links Ravikumar, Deepak Soufleri, Efstathia Hashemi, Abolfazl Roy, Kaushik Machine Learning Artificial Intelligence Cryptography and Security Deep Neural Nets (DNNs) have become a pervasive tool for solving many emerging problems. However, they tend to overfit to and memorize the training set. Memorization is of keen interest since it is closely related to several concepts such as generalization, noisy learning, and privacy. To study memorization, Feldman (2019) proposed a formal score, however its computational requirements limit its practical use. Recent research has shown empirical evidence linking input loss curvature (measured by the trace of the loss Hessian w.r.t inputs) and memorization. It was shown to be ~3 orders of magnitude more efficient than calculating the memorization score. However, there is a lack of theoretical understanding linking memorization with input loss curvature. In this paper, we not only investigate this connection but also extend our analysis to establish theoretical links between differential privacy, memorization, and input loss curvature. First, we derive an upper bound on memorization characterized by both differential privacy and input loss curvature. Second, we present a novel insight showing that input loss curvature is upper-bounded by the differential privacy parameter. Our theoretical findings are further empirically validated using deep models on CIFAR and ImageNet datasets, showing a strong correlation between our theoretical predictions and results observed in practice. |
| title | Unveiling Privacy, Memorization, and Input Curvature Links |
| topic | Machine Learning Artificial Intelligence Cryptography and Security |
| url | https://arxiv.org/abs/2402.18726 |