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].
An Adaptive Test of Independence with Analytic Kernel Embeddings
Authors: Wittawat Jitkrittum, Zoltán Szabó, Arthur Gretton
ICML 2017 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In real-world benchmarks, independence tests using the optimized features perform comparably to the state-of-the-art quadratic-time HSIC test, and outperform competing O(n) and O(n log n) tests. In this section, we empirically study the performance of the proposed method on both toy (Section 3.1) and real problems (Section 3.2). |
| Researcher Affiliation | Academia | 1Gatsby Unit, University College London, UK. 2CMAP, École Polytechnique, France. |
| Pseudocode | No | The paper describes methods in mathematical notation and prose, but does not include any structured pseudocode or algorithm blocks. |
| Open Source Code | Yes | Code is available at https://github.com/wittawatj/fsic-test. |
| Open Datasets | Yes | Million Song Data subset: https://archive.ics. uci.edu/ml/datasets/Year Prediction MSD. Video Story46K dataset: https://ivi.fnwi.uva.nl/isis/mediamill/datasets/videostory.php. |
| Dataset Splits | Yes | For a sample of size n, NFSIC-opt uses half the sample for parameter tuning, and the other disjoint half for the test. |
| Hardware Specification | No | No specific hardware details (such as CPU/GPU models, memory, or cloud instance types) used for running experiments are provided. |
| Software Dependencies | No | The paper does not provide specific version numbers for software dependencies or libraries used in the implementation or experiments. |
| Experiment Setup | Yes | The parameters of NFSIC-opt are σx, σy, and J locations of size (dx + dy)J. We treat all the parameters as a long vector in R2+(dx+dy)J and use gradient ascent to optimize ˆλn/2. ... The regularization parameter γn in NFSIC is fixed to a small value, and is not optimized. ... Gaussian widths σx and σy are set according to the widely used median heuristic... We set J = 10, use 10 inducing points in Ny HSIC, and 10 random Fourier features in FHSIC and RDC. ... The significance level α is set to 0.05. |