Pairwise Learning with Differential Privacy Guarantees
Authors: Mengdi Huai, Di Wang, Chenglin Miao, Jinhui Xu, Aidong Zhang694-701
AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments on real-world datasets are conducted to evaluate the proposed algorithms and the experimental results support our theoretical analysis. |
| Researcher Affiliation | Academia | 1Department of Computer Science, University of Virginia 2Department of Computer Science and Engineering, State University of New York at Buffalo |
| Pseudocode | Yes | Algorithm 1 Online Pairwise Private GIGA-Strongly Convex (On Pair Str C); Algorithm 2 Online Pairwise Private GIGA-Convex (On Pair C); Algorithm 3 Offline Pairwise Private GIGA-Strongly Convex (Off Pair Str C); Algorithm 4 Pairwise Private GIGA-Convex (Off Pair C) |
| Open Source Code | No | The paper does not provide any explicit statements about making the source code available or include links to a code repository. |
| Open Datasets | Yes | We use four real-world datasets that are widely adopted in pairwise learning tasks. These datasets are the Diabetes dataset, the Diabetic Retinopathy dataset, the Hepatitis dataset and the Cancer dataset (Dua and Graff 2017). |
| Dataset Splits | No | The paper mentions varying training size and using a test set, but it does not specify the train/validation/test splits (e.g., percentages, counts, or methodology) needed for reproduction. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware (e.g., GPU/CPU models, memory) used for running the experiments. |
| Software Dependencies | No | The paper does not provide specific software names with version numbers for reproducibility (e.g., Python version, library versions). |
| Experiment Setup | Yes | We add additional ℓ2 regularization λ 2 w 2 2 with λ = 10 3 to loss function for the strongly convex case. ... For the KNN classifier, we set K to be 3. ... In these experiments, the value of δ is fixed as 1 n, and we consider three cases where the parameter ϵ is set to be 0.5, 1.5 and 2.5, respectively. |