Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].
FLIP: A Utility Preserving Privacy Mechanism for Time Series
Authors: Tucker McElroy, Anindya Roy, Gaurab Hore
JMLR 2023 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Numerical studies explore the practical performance of the new method, and an empirical application to labor force data show the method s favorable utility properties in comparison to other competing privacy mechanisms. In this section we present the results of a modest simulation study, as well as some real data analysis. |
| Researcher Affiliation | Collaboration | Tucker Mc Elroy EMAIL Research and Methodology Directorate, U.S. Census Bureau 4600 Silver Hill Road,Washington, D.C. 20233-9100, USA Anindya Roy EMAIL U.S. Census Bureau Department of Mathematics and Statistics University of Maryland, Baltimore County 1000 Hilltop Cir, Baltimore, MD 21250 Gaurab Hore EMAIL Department of Mathematics and Statistics University of Maryland, Baltimore County 1000 Hilltop Cir, Baltimore, MD 21250 |
| Pseudocode | No | The paper includes a section 3.3 titled "Implementation of FLIP" which describes the steps for implementing the mechanism in a numbered list, but it does not present these steps in a formal pseudocode or algorithm block format. |
| Open Source Code | No | The paper does not contain any explicit statements about releasing source code or provide links to a code repository for the described methodology. |
| Open Datasets | Yes | All data were extracted from the QWI Explorer website QWI on 17th March 2022 at 10:45 pm. QWI Explorer, U.S. Census Bureau. https://qwiexplorer.ces.census.gov/. Accessed: 03-17-2022 at 10:45 pm. |
| Dataset Splits | No | For simulated data, the paper states, "We generated 500 Monte Carlo replicates of the bivariate series in each case, with sample size T = 200." For real data, it mentions an "observation period of Q1 1995 through Q4 2019." However, it does not specify any training, validation, or test splits for either the simulated or real datasets. |
| Hardware Specification | No | The paper does not provide specific details about the hardware (e.g., GPU/CPU models, memory, cloud platforms) used for running the experiments or simulations. |
| Software Dependencies | No | The paper does not list specific software dependencies with version numbers for its implementation. While it refers to concepts and tools common in time series analysis, it does not specify the software environment used for the experiments. |
| Experiment Setup | Yes | For implementing FLIP we chose the truncation order for the all-pass filter to be K = 25 for the cepstral representation and to be M = 45 for the filter coefficients. The cross correlation was chosen to be either 0.1 or 0.7. For the present example we chose σ2 = 0.5. The privacy budget δ is chosen to be 0.1. we assume d = 3, i.e. the trend is a polynomial of order three. we set K = 25 and M = 25. |