Causal Identification with Matrix Equations

Authors: Sanghack Lee, Elias Bareinboim

NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical In this paper, we develop a new causal identification algorithm which utilizes both graphical criteria and matrix equations. Specifically, we first characterize the relationships between certain graphically-driven formulae and matrix multiplications. With such characterizations, we broaden the spectrum of proxy variable based identification conditions and further propose novel intermediary criteria based on the pseudoinverse of a matrix. Finally, we devise a causal effect identification algorithm, which accepts as input a collection of marginal, conditional, and interventional distributions, integrating enriched matrix-based criteria into a graphical identification approach. We develop a general identification algorithm that amalgamates graphical and matrical approaches, returning an identification formula for a causal query given a causal diagram and a set of marginal, experimental, and conditional distributions. We show that this method subsumes current state of the art in the literature. Theorem 4. ID-ME is sound. Theorem 5. ID-ME strictly subsumes proxy criteria [8, 15], GID(-PO) [14, 12], m ID, or e ID [11].
Researcher Affiliation Academia Sanghack Lee Graduate School of Data Science Seoul National University Seoul, South Korea sanghack@snu.ac.kr Elias Bareinboim Department of Computer Science Columbia University New York, USA eb@cs.columbia.edu
Pseudocode Yes We present a causal identification algorithm ID-ME (Alg. 1) and provides "Algorithm 1 ID-ME".
Open Source Code No The paper does not provide an explicit statement about releasing source code for the described methodology, nor does it provide a link to a code repository. It only refers to a technical report for omitted proofs and derivations.
Open Datasets No The paper is theoretical and does not involve the use of any datasets for training or evaluation.
Dataset Splits No The paper is theoretical and does not involve the use of datasets or their partitioning into train/validation/test splits.
Hardware Specification No The paper is theoretical and does not describe any hardware specifications used for experiments.
Software Dependencies No The paper is theoretical and does not mention specific software dependencies with version numbers required for replication.
Experiment Setup No The paper is theoretical and does not include details about an experimental setup, hyperparameters, or training settings.