Dynamical Systems as Temporal Feature Spaces
Authors: Peter Tino
JMLR 2020 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In section 8, 'Illustrative examples', the paper states: 'In all illustrations we will use state space dimensionality N = 100 and re-normalise the dynamic coupling W R100 100 to largest singular value ν = 0.995. The input coupling w is renormalized to unit length. The past horizon will be set to τ = 200. We will show motifs with motif weights up to 10 2 of the highest motif weight.' It also presents figures showing 'Fourier coefficient distribution' and 'Relative area covered by Fourier coefficients' which involve quantitative analysis of computational simulations. |
| Researcher Affiliation | Academia | Peter Tino EMAIL School of Computer Science University of Birmingham Birmingham, B15 2TT, UK |
| Pseudocode | No | The paper describes mathematical derivations and theoretical concepts, along with illustrative examples. It does not contain any clearly labeled pseudocode or algorithm blocks. |
| Open Source Code | No | The paper states: 'License: CC-BY 4.0, see https://creativecommons.org/licenses/by/4.0/. Attribution requirements are provided at http://jmlr.org/papers/v21/19-589.html.' This refers to the license for the paper itself, not an explicit statement about the release of source code for the methodology described in the paper. No links to code repositories or statements about supplementary code were found. |
| Open Datasets | No | The paper does not use or refer to any publicly available datasets for its illustrative examples. Instead, it generates random matrices and configurations (e.g., 'generated 100 realisations of f W with elements Wi,j randomly distributed according to the standard normal distribution N(0, 1)') for its analysis, which are internal to the simulation. |
| Dataset Splits | No | The paper conducts theoretical analysis and computational illustrations involving generated data rather than experiments on predefined datasets requiring training/test/validation splits. Therefore, there is no mention of dataset splits. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware used to run the computational illustrations or analyses. |
| Software Dependencies | No | The paper does not mention any specific software, libraries, or their version numbers used for its analysis or computational illustrations. |
| Experiment Setup | Yes | In section 8, 'Illustrative examples', the paper specifies key parameters for its computational analysis: 'In all illustrations we will use state space dimensionality N = 100 and re-normalise the dynamic coupling W R100 100 to largest singular value ν = 0.995. The input coupling w is renormalized to unit length. The past horizon will be set to τ = 200. We will show motifs with motif weights up to 10 2 of the highest motif weight.' It also details the generation of W (e.g., 'elements Wi,j were generated i.i.d. from normal distribution N(0, 1)') and w (e.g., 'w generated randomly i.i.d. from N(0, 1)' or 'a vector of ones with signs prescribed by the first N = 100 digits of binary expansion of π'). |