Calibrated Reliable Regression using Maximum Mean Discrepancy
Authors: Peng Cui, Wenbo Hu, Jun Zhu
NeurIPS 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments on non-trivial real datasets show that our method can produce well-calibrated and sharp prediction intervals, which outperforms the related state-of-the-art methods. |
| Researcher Affiliation | Collaboration | Peng Cui1 2, Wenbo Hu1 2, Jun Zhu1 1 Dept. of Comp. Sci. & Tech., Institute for AI, BNRist Center Tsinghua-Bosch Joint ML Center, THBI Lab, Tsinghua University, Beijing, 100084 China 2 Real AI |
| Pseudocode | Yes | Algorithm 1 Deep calibrated reliable regression model. |
| Open Source Code | No | The paper does not contain any statements about releasing source code for the described methodology or provide a link to a code repository. |
| Open Datasets | Yes | We use several public datasets from UCI repository [2] and Kaggle [1]: 1) for the timeseries task: Pickups, Bike-sharing, PM2.5, Metro-traffic and Quality; 2) for the regression task: Power Plant, Protein Structure, Naval Propulsion and wine. |
| Dataset Splits | No | The paper specifies training and testing splits (e.g., “70% training data and 30% test data” or “80% of each data-set for training and the rest for testing”) but does not mention a separate validation split or how it was handled. |
| Hardware Specification | Yes | on the wine dataset on GTX1080Ti. |
| Software Dependencies | No | The paper does not specify any software dependencies (e.g., libraries, frameworks) with version numbers required to replicate the experiments. |
| Experiment Setup | Yes | For time-series forecasting tasks, we construct an LSTM model with two hidden layers (128 hidden units and 64 units respectively) and a linear layer for making the final predictions. ... The details of hyperparameters setting can be found in Appendix B.2. |