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 Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The Shift Toward Open and Reproducible AI Research," Under Review, 2026. doi: 10.48550/arXiv.2606.16974..
Robust Inverse Covariance Estimation under Noisy Measurements
Authors: Jun-Kun Wang, Shou-de Lin
ICML 2014 | Venue PDF | LLM Run Details | Input Tokens: 17,133 Total number of tokens sent to the LLM as input for this paper's analysis. | Output Tokens: 4,634 Total number of tokens produced by the LLM (including reasoning/thinking tokens) for this paper's analysis.
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments on time series prediction and classification under noisy condition demonstrate the effectiveness of the approach. |
| Researcher Affiliation | Collaboration | Jun-Kun Wang EMAIL Intel-NTU, National Taiwan University, Taiwan Shou-de Lin EMAIL Intel-NTU, National Taiwan University, Taiwan |
| Pseudocode | Yes | Algorithm 1 Adjusting the inverse covariance that guarantees positive semi-definiteness |
| Open Source Code | Yes | The codes to reproduce the experiments are available on the first author s page https://sites.google.com/site/wangjim123. |
| Open Datasets | Yes | 1) Stock: The data are downloaded from Yahoo Finance... 2) Temperature (medium variable size): The data are downloaded from National Oceanic and Atmospheric Administration (NOAA) 1... 3) Temperature (large variable size)... four datasets, all are available on http://www.csie.ntu.edu.tw/~cjlin/libsvmtools/datasets/. |
| Dataset Splits | Yes | The last 30 trading days are reserved for testing; the second to last 30 days are for validation; and the remaining data are for training. ... For each dataset, we random split data 5 times that 80 percent of data are for cross-validation and the remaining for testing. |
| Hardware Specification | Yes | Our experiment is run on a machine with dual core 2.66 GHz (INTEL E5500) and 32GB memory. |
| Software Dependencies | No | The paper does not provide specific version numbers for any software dependencies or libraries used. |
| Experiment Setup | Yes | We use grid search to tune the regularization parameters. For our method, denote c as a vector whose entries are perturbation bound cg, the regularization vector is searched by c times [10 8, 10 7, . . . , 102] over the grid. ... input features contain previous three historic values... We compare our method with 1) the work of Hsieh (2011)... and 2) the work of Yuan (2010)... |