Improving Multi-Step Prediction of Learned Time Series Models

Authors: Arun Venkatraman, Martial Hebert, J.. Bagnell

AAAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We present experimental results of our method, DAD, and show significant improvement over the traditional approach in two notably different domains, dynamic system modeling and video texture prediction.
Researcher Affiliation Academia Arun Venkatraman, Martial Hebert, and J. Andrew Bagnell Robotics Institute, Carnegie Mellon University, Pittsburgh, PA 15213 arunvenk@cs.cmu.edu, {hebert, dbagnell}@ri.cmu.edu
Pseudocode Yes Algorithm 1 DATA AS DEMONSTRATOR (DAD)
Open Source Code No The paper does not provide concrete access to source code for the methodology described.
Open Datasets Yes The final dynamical system we test on is the simulated helicopter from (Abbeel and Ng 2005a).
Dataset Splits No Algorithm 1 mentions 'return Mn with lowest error on validation trajectories' but the experimental setup does not specify the size or split method for validation data.
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, memory amounts) used for running its experiments.
Software Dependencies No The paper mentions 'Random Fourier Feature Regression' and 'Gaussian kernel' but does not provide specific version numbers for any software components.
Experiment Setup Yes In the experiments described below, we choose m = 500 and use the Gaussian kernel. The hyperparameters λ and kernel bandwidth σ were chosen using cross-validated grid search.