Differentially Private Ordinary Least Squares
Authors: Or Sheffet
ICML 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | 6. Experiment: t-Values of Output. Goal. We set to experiment with the outputs of Algorithms 1 and 2... We tested both algorithms in two settings. The first is over synthetic data... The second setting is over real-life data... We repeated each algorithm 100 times. Results. We plot the t-values we get from Algorithms 1 and 2... |
| Researcher Affiliation | Academia | 1Computing Science Dept., University of Alberta, Edmonton AB, Canada. This work was done when the author was at Harvard University, supported by NSF grant CNS-123723. Correspondence to: Or Sheffet <osheffet@ualberta.ca>. |
| Pseudocode | Yes | Algorithm 1 Outputting a private Johnson-Lindenstrauss projection of a matrix. |
| Open Source Code | No | The paper does not provide concrete access to source code. It does not contain an explicit statement that the authors are releasing their code for the work described in this paper, nor does it provide a direct link to a code repository. |
| Open Datasets | Yes | The second setting is over real-life data. We ran the two algorithms over diabetes dataset collected over ten years (1999-2008) taken from the UCI repository (Strack et al., 2014). |
| Dataset Splits | No | The paper does not provide specific dataset split information (exact percentages, sample counts, or detailed splitting methodology) for training, validation, or test sets. It mentions feeding the data in varying sizes but not explicit splits. |
| Hardware Specification | No | The paper does not provide specific hardware details (exact GPU/CPU models, processor types with speeds, memory amounts, or detailed computer specifications) used for running its experiments. |
| Software Dependencies | No | The paper does not provide specific ancillary software details, such as library or solver names with version numbers, needed to replicate the experiment. |
| Experiment Setup | Yes | We set ϵ = 0.25 and δ = 10 6. We chose β = (0.5, 0.25, 0) so the first coordinate is twice as big a the second but of opposite sign, and moreover, y is independent of the 3rd feature. The variance of the label is also set to 1, and so the variance of the homosedastic noise equals to σ2 = 1 (0.5)2 ( 0.25)2. The number of observations n ranges from n = 1000 to n = 100000. |