Bayesian Posterior Approximation via Greedy Particle Optimization
Authors: Futoshi Futami, Zhenghang Cui, Issei Sato, Masashi Sugiyama3606-3613
AAAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Numerical experiments We experimentally confirmed the usefulness of the proposed method compared with SVGD and SP in both toy datasets and real world datasets. Other than comparing the performance measured in terms of the accuracy or RMSE of the proposed method with SVGD and SP, we also have the following two purposes for the experiments. |
| Researcher Affiliation | Academia | 1The Univiersity of Tokyo 2RIKEN |
| Pseudocode | Yes | Algorithm 1: Frank-Wolfe (FW) Algorithm |
| Open Source Code | No | The paper does not provide an explicit statement or link for the open-sourcing of the described methodology's code. A link to an arXiv preprint is provided, but this is not a code repository. |
| Open Datasets | Yes | As the dataset, we used Covertype (Dheeru and Karra Taniskidou 2017), with 581,012 data points and 54 features. We used the Naval data from the UCI (Dheeru and Karra Taniskidou 2017), which contains 11,934 data points and 17 features. |
| Dataset Splits | No | The paper states 'we split dataset 90% for training and 10% for testing', but does not explicitly mention a validation split or its details. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware (e.g., CPU, GPU models, memory) used for running the experiments. |
| Software Dependencies | No | The paper mentions 'Adam (Kingma and Ba 2014)' as an optimizer used, but does not specify any version numbers for Adam or any other software dependencies like Python, PyTorch, or TensorFlow. |
| Experiment Setup | Yes | We used a neural network with one hidden layer, 50 units, and the Re LU activation function. As the dataset, we used the Naval data from the UCI (Dheeru and Karra Taniskidou 2017), which contains 11,934 data points and 17 features. The posterior dimension was 953. ... where α = 1.0 and β = 0.5 are used as suggested in the original paper. |