Asynchronous Distributed Gaussian Process Regression
Authors: Zewen Yang, Xiaobing Dai, Sandra Hirche
AAAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Numerical simulations within GPgym across regression tasks with real-world data sets and dynamical control scenarios demonstrate the effectiveness and applicability of Async DGP. |
| Researcher Affiliation | Academia | Technical University of Munich EMAIL, EMAIL, EMAIL |
| Pseudocode | Yes | Algorithm 1: Asyn GP |
| Open Source Code | Yes | The GPgym platform, including data set, code, and instructions, is provided in (Dai and Yang 2024). |
| Open Datasets | Yes | The GPgym platform, including data set, code, and instructions, is provided in (Dai and Yang 2024). Regression Benchmark The regression performance on the three datasets, namely KIN40K (8-dimensional input, 10K data), SARCOS (21dimensional input, 44484 data), and PUMADYN32NM (32dimensional input, 7168 data) are evaluated. |
| Dataset Splits | No | The paper mentions that each GP model receives streaming data after the prediction process is finished, leading to variations in their training data. However, it does not provide specific train/test/validation splits or percentages for the datasets used in the benchmarks. |
| Hardware Specification | No | The paper mentions the distributed system is "connected through UDP via Wi-Fi" but does not specify any particular hardware models (e.g., GPU, CPU models, memory) used for running the experiments. |
| Software Dependencies | No | The paper mentions using "MATLAB simulation platform" for GPgym and Lo G-GP for online learning, but no specific version numbers are provided for any software components. |
| Experiment Setup | Yes | The standardized mean squared errors (SMSE) for regression, with the information set threshold I set to 4 and 10, are shown in Figure 3, respectively. By employing the proposed learning-based control defined in (21) using the DGP composed of 4 models with I = 20... The control gains are set to λ1 = 2 and λ2 = 10 and I = 20. |