Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].

Identifiability and Asymptotics in Learning Homogeneous Linear ODE Systems from Discrete Observations

Authors: Yuanyuan Wang, Wei Huang, Mingming Gong, Xi Geng, Tongliang Liu, Kun Zhang, Dacheng Tao

JMLR 2024 | Venue PDF | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We also construct simulations with various system dimensions to illustrate the established theoretical results. Keywords: identifiability, linear ODEs, asymptotic analysis, nonlinear least squares, causal discovery. 4. Simulations In this section, we illustrate the theoretical results established in Section 3 by simulation.
Researcher Affiliation Academia School of Mathematics and Statistics University of Melbourne, Melbourne, Australia; School of Computer Science, Faculty of Engineering The University of Sydney, Sydney, Australia; Carnegie Mellon University, Pittsburgh, PA, USA; Mohamed bin Zayed University of Artificial Intelligence, Abu Dhabi, UAE
Pseudocode No The paper does not contain any structured pseudocode or algorithm blocks.
Open Source Code No The paper does not explicitly state that source code for the methodology is provided or publicly available.
Open Datasets No For each d = 2, 3, 4, we first randomly generate a d d parameter matrix A d and a d 1 initial condition x 0d as the true system parameters for each d-dimensional ODE system (1). This indicates the data was simulated, not obtained from an open dataset.
Dataset Splits No Then n equally-spaced noisy observations are generated based on Equation (4) in [0, 1] time interval with error term ϵi N(0, diag(0.052, . . . , 0.052)). We tested various sample sizes for each d-dimensional ODE system. For each configuration, we run 200 random replications. The paper simulates data for each experiment run and does not use pre-defined training/test/validation splits from a larger fixed dataset.
Hardware Specification No The paper does not provide specific hardware details used for running its experiments.
Software Dependencies No The paper mentions using a 'bound-constrained minimization technique (Branch et al., 1999)' but does not specify any software names with version numbers for implementation.
Experiment Setup Yes For each d = 2, 3, 4, we first randomly generate a d d parameter matrix A d and a d 1 initial condition x 0d as the true system parameters for each d-dimensional ODE system (1). Without loss of generality, we set T = 1. Then n equally-spaced noisy observations are generated based on Equation (4) in [0, 1] time interval with error term ϵi N(0, diag(0.052, . . . , 0.052)). We tested various sample sizes for each d-dimensional ODE system. For each configuration, we run 200 random replications. Firstly, we initialize the parameter with a value close to the true parameter (for example, θ ± 0.001). Secondly, we constrain the bounds of the parameter within a reasonable neighbourhood of the true parameter (for example, [θ − 0.5, θ + 0.5]).