LSH-SMILE: Locality Sensitive Hashing Accelerated Simulation and Learning

Authors: Chonghao Sima, Yexiang Xue

NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimentally, we demonstrate that LSH-SMILE simulates physics systems at comparable quality with exact approaches, but with way less time and space complexity.
Researcher Affiliation Academia Chonghao Sima Department of Computer Science Purdue University West Lafayette, IN, USA, 47907 simac@purdue.edu Yexiang Xue Department of Computer Science Purdue University West Lafayette, IN, USA, 47907 yexiang@purdue.edu
Pseudocode Yes The pseudo code of LSH-SMILE for forward simulation is in Figure 4.
Open Source Code Yes Did you include the code, data, and instructions needed to reproduce the main experimental results (either in the supplemental material or as a URL)? [Yes] they will be provided in the supplementary material
Open Datasets No The dataset is synthetic and contains 1700 frames of grain growth ground truth simulation results. ... We use our own synthetic data
Dataset Splits No The paper mentions training epochs but does not explicitly detail train/validation/test dataset splits within the main text. The checklist states 'they will be provided in the supplementary material' but this is not in the main paper content.
Hardware Specification No The paper provides running times and memory usage in Table 1, but does not specify the exact hardware components (e.g., CPU, GPU models, or memory capacity) used for the experiments in the main text.
Software Dependencies No The paper mentions using the 'Torch framework' but does not specify its version number or other software dependencies with version numbers.
Experiment Setup Yes We set dt to 0.05 in this simulation. The image size is 128 by 128. For LSH parameter, r and r0 are set to be 0.01, K is 3 and L is 10. ... We set dt to 0.1 in this simulation. The image size is 128 by 128. For LSH parameter, r is set to be 0.0001, K is 3 and L is 10. ... LSH-SMILE uses stochastic gradient descent, while the baseline use the Adam optimizer in our experiment.