Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Causal Learning in Biomedical Applications: A Benchmark

Petr Ryšavý, Xiaoyu He, Jakub Mareček

TL;DR

A benchmark for methods in causal learning using time series using time series is presented, based on a real-world scenario of the Krebs cycle that is used in cells to release energy.

Abstract

Learning causal relationships between a set of variables is a challenging problem in computer science. Many existing artificial benchmark datasets are based on sampling from causal models and thus contain residual information that the ${R} ^2$-sortability can identify. Here, we present a benchmark for methods in causal learning using time series. The presented dataset is not ${R}^2$-sortable and is based on a real-world scenario of the Krebs cycle that is used in cells to release energy. We provide four scenarios of learning, including short and long time series, and provide guidance so that testing is unified between possible users.

Causal Learning in Biomedical Applications: A Benchmark

TL;DR

A benchmark for methods in causal learning using time series using time series is presented, based on a real-world scenario of the Krebs cycle that is used in cells to release energy.

Abstract

Learning causal relationships between a set of variables is a challenging problem in computer science. Many existing artificial benchmark datasets are based on sampling from causal models and thus contain residual information that the -sortability can identify. Here, we present a benchmark for methods in causal learning using time series. The presented dataset is not -sortable and is based on a real-world scenario of the Krebs cycle that is used in cells to release energy. We provide four scenarios of learning, including short and long time series, and provide guidance so that testing is unified between possible users.
Paper Structure (24 sections, 3 equations, 4 figures, 3 tables)

This paper contains 24 sections, 3 equations, 4 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Preliminaries and Related Work
  3. System Identification and Linear Dynamic Systems (LDS)
  4. Linear Additive Noise Models
  5. Bayesian networks
  6. Counterfactual Framework
  7. Granger Causality
  8. Instrumental Variables
  9. Tractable Probabilistic Models
  10. The Challenge
  11. Cycles
  12. Hidden state and Mixture-models
  13. Model learning
  14. Non-linear dependencies
  15. The Benchmark
  16. ...and 9 more sections

Figures (4)

  • Figure 1: Illustration of the Kreb's cycle reactions used to simulate the concentrations.
  • Figure 2: An illustration of the adjacency matrix produced by various methods and the ground truth matrix representing the set of reactions. Black squares represent $1$ an edge in the adjacency matrix, grey $0$.
  • Figure 3: An illustration of the F1-score of various methods on the Krebs dataset. Please note that the implementation of DyNoTears in CausalNex is deterministic, thus providing the same result each time. To calculate the error bars, randomly selected $10\,\%$ of the data were put aside, and then results were averaged over $10$ repeats of this procedure.
  • Figure 4: An illustration of the time requirements of various methods on the Krebs dataset. The error bars show the standard deviation of the measurements calculated from $10$ repeats.