Learning Why Things Change: The Difference-Based Causality Learner

By Mark Voortman, Denver Dash, Marek J Druzdzel. 2010

42 citations recorded by the source corpus.

In this paper, we present the Difference-Based Causality Learner (DBCL), an algorithm for learning a class of discrete-time dynamic models that represents all causation across time by means of difference equations driving change in a system. We motivate this representation with real-world mechanical systems and prove DBCL's correctness for learning structure from time series data, an endeavour that is complicated by the existence of latent derivatives that have to be detected. We also prove that, under common assumptions for causal discovery, DBCL will identify the presence or absence of feedback loops, making the model more useful for predicting the effects of manipulating variables when the system is in equilibrium. We argue analytically and show empirically the advantages of DBCL over vector autoregression (VAR) and Granger causality models as well as modified forms of Bayesian and constraint-based structure discovery algorithms. Finally, we show that our algorithm can discover causal directions of alpha rhythms in human brains from EEG data.

Subfields: Artificial Intelligence

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Learning Causal Structure from Undersampled Time Series (2013, methodology): “…" experimental strategy is often to provide some exogenous shock to the dynamical system, measure its evolution, and then apply a causal learning algorithm to the resulting time series data, such as Demiralp and Hoover (2003); Moneta et al. (2011); Friedman et al. (1999); Voortman et al. (2010).”
Mesochronal Structure Learning (2015, methodology): “There are multiple algorithms for learning graphical structure from time series data [6, 7, 11, 14, 15], but they all assume that τ M is at least as fast as τ S .”
From Random Differential Equations to Structural Causal Models: the stochastic case (2018, background): “Other related work on discrete-time dynamical system and causality which does not require a single static equilibrium assumption is (Voortman et al., 2010).”
Rate-Agnostic (Causal) Structure Learning (2015, background): “There has been substantial work on modeling the statistics of time series, but relatively less on learning causal structure, and almost all of that assumes that the measurement and causal timescales match [1–5].”