Deep Learning for Precipitation Nowcasting: A Benchmark and A New Model
Authors: Xingjian Shi, Zhihan Gao, Leonard Lausen, Hao Wang, Dit-Yan Yeung, Wai-kin Wong, Wang-chun WOO
NeurIPS 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We train all models using the Adam optimizer [14] with learning rate equal to 10 4 and momentum equal to 0.5. For the RNN models, we use the encoding-forecasting structure introduced previously with three RNN layers. All RNNs are either Conv GRU or Traj GRU and all use the same set of hyperparameters. The HKO-7 dataset used in the benchmark contains radar echo data from 2009 to 2015 collected by HKO. The overall evaluation results are summarized in Table 3. |
| Researcher Affiliation | Collaboration | Xingjian Shi, Zhihan Gao, Leonard Lausen, Hao Wang, Dit-Yan Yeung Department of Computer Science and Engineering Hong Kong University of Science and Technology {xshiab,zgaoag,lelausen,hwangaz,dyyeung}@cse.ust.hk Wai-kin Wong, Wang-chun Woo Hong Kong Observatory Hong Kong, China {wkwong,wcwoo}@hko.gov.hk |
| Pseudocode | No | The paper does not contain any structured pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not provide an explicit statement or link to the open-source code for the described methodology. |
| Open Datasets | No | The paper introduces the 'HKO-7 dataset' and describes its characteristics, but it does not provide concrete access information (e.g., URL, DOI, or a formal citation for public access) for the dataset itself. |
| Dataset Splits | Yes | As rainfall events occur sparsely, we select the rainy days based on the rain barrel information to form our final dataset, which has 812 days for training, 50 days for validation and 131 days for testing. |
| Hardware Specification | No | The paper does not specify any particular hardware components (e.g., GPU or CPU models, memory details) used for conducting the experiments. |
| Software Dependencies | No | The paper mentions using the 'Adam optimizer' and 'Ada Grad' but does not provide specific version numbers for these or any other software libraries or frameworks. |
| Experiment Setup | Yes | We train all models using the Adam optimizer [14] with learning rate equal to 10 4 and momentum equal to 0.5. For the RNN models, we use the encoding-forecasting structure introduced previously with three RNN layers. All RNNs are either Conv GRU or Traj GRU and all use the same set of hyperparameters. For the RNN models, we train them for 200,000 iterations with norm clipping threshold equal to 1 and batch size equal to 4. For the CNN models, we train them for 100,000 iterations with norm clipping threshold equal to 50 and batch size equal to 32. The strides of the middle downsampling and upsampling layers are chosen to be 2. The numbers of filters for the three RNNs are 64, 96, 96 respectively. For Traj GRU and Conv GRU models, we use a 3-layer encoding-forecasting structure with the number of filters for the RNNs set to 64, 192, 192. We use kernel size equal to 5 5, 5 5, 3 3 for the Conv GRU models while the number of links is set to 13, 13, 9 for the Traj GRU model. |