Gaussian quadrature for matrix inverse forms with applications
Authors: Chengtao Li, Suvrit Sra, Stefanie Jegelka
ICML 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We illustrate empirical consequences of our results by using quadrature to accelerate machine learning tasks involving determinantal point processes and submodular optimization, and observe tremendous speedups in several instances. |
| Researcher Affiliation | Academia | Chengtao Li CTLI@MIT.EDU Suvrit Sra SUVRIT@MIT.EDU Stefanie Jegelka STEFJE@MIT.EDU Massachusetts Institute of Technology |
| Pseudocode | Yes | Algorithm 1 Gauss Quadrature Lanczos (GQL) ... Algorithm 2 Efficient Retrospective Framework ... Algorithm 3 Gauss-DPP (L) ... Algorithm 4 DPPJUDGE |
| Open Source Code | No | The paper does not provide any statement about making its own source code publicly available. |
| Open Datasets | Yes | The first two of small/medium-sized datasets, Abalone and Wine2, are popular datasets for regression... Available at http://archive.ics.uci.edu/ml/. ... The final two large datasets datasets are Epinions (Who-trustswhom network of Epinions) and Slashdot (Slashdot social network from Feb. 2009) 3 ... Available at https://snap.stanford.edu/data/. |
| Dataset Splits | No | The paper does not provide specific details on training, validation, or test dataset splits, such as percentages or sample counts. While it uses existing datasets, it doesn't describe how they were partitioned for the experiments. |
| Hardware Specification | No | The paper does not specify the hardware used for running its experiments, such as particular CPU or GPU models. |
| Software Dependencies | No | The paper mentions general software usage but does not provide specific version numbers for any key software components or libraries (e.g., 'Python 3.x', 'PyTorch 1.x'). |
| Experiment Setup | Yes | We set the bandwidth parameter for Abalone as σ = 0.15 and that for Wine as σ = 1 and the cut-off parameter as 3σ for both datasets, as in (Gittens & Mahoney, 2013). ... For all datasets we add an 1E-3 times identity matrix to ensure positive definiteness. ... DPP is initialized with random subsets of size N/3 and corresponding running times are averaged over 1,000 iterations of the chain. All results are averaged over 3 runs. |