Summary of Impact Of Privacy Parameters on Deep Learning Models For Image Classification, by Basanta Chaulagain
Impact of Privacy Parameters on Deep Learning Models for Image Classification
by Basanta Chaulagain
First submitted to arxiv on: 9 Dec 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Cryptography and Security (cs.CR)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The project aims to develop differentially private deep learning models for image classification on CIFAR-10 datasets and analyze the impact of various privacy parameters on model accuracy. It explores five different deep learning architectures: ConvNet, ResNet18, EfficientNet, ViT, and DenseNet121. The researchers implemented three supervised classifiers (K-Nearest Neighbors, Naive Bayes Classifier, and Support Vector Machine) and evaluated their performance under varying settings. The best-performing model to date is EfficientNet with a test accuracy of 59.63%. The study uses the Adam optimizer, batch size 256, epoch size 100, epsilon value 5.0, learning rate 1e-3, clipping threshold 1.0, and noise multiplier 0.912. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The project develops special computer models that keep your pictures private. It tests five different models to see which one works best. The models are trained on a dataset of images and then tested to see how well they can recognize objects. The best model is called EfficientNet, and it gets 59.63% of the test questions right. |
Keywords
» Artificial intelligence » Deep learning » Image classification » Naive bayes » Supervised » Support vector machine » Vit
