Causal Identification for Complex Functional Longitudinal Studies
Authors: Andrew Ying
ICLR 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We examine our identification framework through Monte Carlo simulations. Our approach addresses significant gaps in current methodologies, providing a solution for functional longitudinal data and paving the way for future estimation work in this domain. In this section, we employ Monte Carlo simulations to empirically assess how the identification works. |
| Researcher Affiliation | -1 | The paper lists "Andrew Ying Irvine, CA 92606, USA EMAIL". The email domain (gmail.com) is personal, and no institutional or company name is provided alongside the author's name. Therefore, there is insufficient information to classify the author's affiliation as academic, industry, or collaborative. |
| Pseudocode | No | The paper includes mathematical decompositions and formulas (e.g., equations 2-5, 6-8 in Section 3.2, and Section A Formal Formulas in the appendix), but these are presented as mathematical expressions, not as structured pseudocode or algorithm blocks. |
| Open Source Code | No | The paper describes Monte Carlo simulations in Section 4 and Appendix D but does not provide any explicit statement about releasing source code for the described methodology or a link to a code repository. |
| Open Datasets | Yes | For example, the Medical Information Mart for Intensive Care IV (MIMIC-IV) (Johnson et al., 2023) is a freely accessible electronic health record (EHR) database that records ICU care data, including physiological measurements, laboratory values, medication administration, and clinical events. |
| Dataset Splits | No | The paper uses Monte Carlo simulations where data is generated according to a specified process (Section 4, Step 3). It describes varying grid sizes and sample sizes for the simulations, but it does not specify training/test/validation splits for any pre-existing dataset. |
| Hardware Specification | No | The paper details Monte Carlo simulations but does not provide any specific hardware details such as GPU models, CPU types, or memory specifications used to run these experiments. |
| Software Dependencies | No | The paper describes mathematical frameworks and Monte Carlo simulations but does not mention specific software dependencies, libraries, or their version numbers that would be required to replicate the experimental results. |
| Experiment Setup | Yes | Step 1: To sharp the focus and ease the computation, we consider a simple setting where there is no mortality or censoring (T = C ), or other measured confounding process, except for the outcome process itself. ... Gaussian processes are particularly well-suited for modeling such smooth and continuous temporal processes. For t [0, 1], consider a potential outcome process Y a(t) capturing potential logarithm of glucose levels, following a Gaussian process with mean process as E(Y a(t)) = a(t), and covariance process as Cov[Y a(t), Y a(s)] = e 3|t s|, t, s [0, 1]. |