Stein Variational Gradient Descent Without Gradient
Authors: Jun Han, Qiang Liu
ICML 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We test our proposed algorithms on both synthetic and real-world examples. |
| Researcher Affiliation | Academia | 1Computer Science, Dartmouth College 2Computer Science, The University of Texas at Austin. |
| Pseudocode | Yes | Algorithm 1 Gradient-Free SVGD (GF-SVGD); Algorithm 2 Annealed SVGD (A-SVGD); Algorithm 3 Annealed GF-SVGD (AGF-SVGD) |
| Open Source Code | No | The paper does not provide any explicit statements or links about the availability of its source code. |
| Open Datasets | Yes | We test the algorithms on Glass dataset and SUSY dataset in Figure 4 from UCI repository (Asuncion & Newman, 2007) for which the dimension of θ is d = 9 and d = 18, respectively. |
| Dataset Splits | No | The paper does not specify any training, validation, or test splits with percentages, absolute counts, or references to predefined splits. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware used for running experiments. |
| Software Dependencies | No | The paper mentions 'Adam optimizer' and 'RBF kernel' but does not specify any software names with version numbers for dependencies needed to replicate the experiment. |
| Experiment Setup | Yes | We use RBF kernel k(x, x ) = exp( x x 2/h) for the updates of our proposed algorithms and the kernel approximation in (24); the bandwidth h is taken to be h=med2/(2 log(n + 1)) where med is the median of the current n particles. When maximum mean discrepancy (MMD) (Gretton et al., 2012) is applied to evaluate the sample quality, RBF kernel is used and the bandwidth is chosen based on the median distance of the exact samples so that all methods use the same bandwidth for a fair comparison. Adam optimizer (Kingma & Ba, 2015) is applied to our proposed algorithms for accelerating convergence. |