Universality of Real Minimal Complexity Reservoir
Authors: Robert Simon Fong, Boyu Li, Peter Tino
AAAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Our findings are supported by empirical studies on real-world time series datasets. We conclude the paper with numerical experiments that validate the structural approximation properties demonstrated in our theoretical analysis. Univariate forecasting performance of the initial and the approximating reservoir systems are compared on two popular datasets used in recent time series forecasting studies (e.g. (Zhou et al. 2020) (adopting their training/validation/test data splits)): ETT The Electricity Transformer Temperature dataset [...] ECL The Electricity Consuming Load. |
| Researcher Affiliation | Academia | 1School of Computer Science, University of Birmingham, Birmingham, B15 2TT, UK 2Department of Mathematical Sciences, New Mexico State University, Las Cruces, New Mexico, 88003, USA EMAIL, EMAIL, EMAIL. |
| Pseudocode | No | The paper includes several mathematical definitions, theorems, and proofs (e.g., Definition 1, Theorem 6, Proof sections), but it does not contain any explicit pseudocode blocks or algorithm listings. |
| Open Source Code | Yes | The source code and data of the numerical analysis is openly available at: Code https://github.com/Lampertos/RSCR |
| Open Datasets | Yes | ETT The Electricity Transformer Temperature dataset [...] Datasets https://github.com/zhouhaoyi/ETDataset ECL The Electricity Consuming Load [...] Datasets https://archive.ics.uci.edu/dataset/321/ electricityloaddiagrams20112014 |
| Dataset Splits | Yes | In this paper we used oil temperature of the ETTm2 dataset for univariate prediction with train/validation/test split being 12/4/4 months. [...] In this paper we used client MT 320 for univariate prediction. The train/validation/test split is 15/3/4 months. |
| Hardware Specification | Yes | For reproducibility of the experiments, all experiments are CPU-based and are performed on Apple M3 Max with 128GB of RAM. |
| Software Dependencies | No | The paper mentions that experiments are "CPU-based" and performed on "Apple M3 Max with 128GB of RAM," but it does not specify any software names with version numbers, such as programming languages, libraries, or frameworks used. |
| Experiment Setup | Yes | In all simulations, we maintain a spectral radius λ = 0.9 and prediction horizon is set to be 300. [...] The readout h of the original reservoir system is trained using ridge regression with a ridge coefficient of 10 9. |