Learning Physics Constrained Dynamics Using Autoencoders

Authors: Tsung-Yen Yang, Justinian Rosca, Karthik Narasimhan, Peter J Ramadge

NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In our experiments on three visual and one sensor measurement tasks, our model imposes interpretability on latent states and achieves improved generalization performance for long-term prediction of system dynamics over state-of-the-art baselines.
Researcher Affiliation Collaboration 1Princeton University 2Siemens Corporation, Corporate Technology
Pseudocode No The paper describes the network architecture but does not include any explicitly labeled pseudocode or algorithm blocks.
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] See Section D in the supplementary material.
Open Datasets No We generate three visual datasets: pendulum, mass-spring-damper (MSD), and a two-body system to compare the performance of ALPS to that of the baseline in the literature. ... In total, we generate 500 training and 500 test trajectories, resulting in 13,000 training and test sequences. ... We use full-scale dynamics of the train wheel suspension system with 18 state elements, 12 observations, 16 input excitations, and 24 system parameters. ... The dataset contains 20 trajectories with 500 steps sampled by 100Hz. We use a 50-50 split to get the training and test datasets with τ = 100.
Dataset Splits No The paper specifies training and test splits but does not explicitly mention or detail a validation set split for either the visual or time-series datasets within the main text.
Hardware Specification No The paper states, 'Did you include the total amount of compute and the type of resources used (e.g., type of GPUs, internal cluster, or cloud provider)? [Yes] See Section D in the supplementary material.' However, specific hardware details such as GPU/CPU models are not provided in the main paper.
Software Dependencies No The paper does not provide specific software names with version numbers for its dependencies. It only mentions 'Matlab [55]' in the context of conventional system identification tools, not as a dependency for their own implementation.
Experiment Setup Yes The sampling rate is 20Hz in the pendulum, 100Hz in MSD, and 6Hz in two-body systems. The observation length τ is 100. ... In practice, we tune the weight for each loss term to accommodate different scales.