Noname manuscript No. (will be inserted by the editor) Replacing Causal Faithfulness with Algorithmic Independence of Conditionals.

Subfields: Causal Inference & Discovery

Citations in the corpus

• A theory of independent mechanisms for extrapolation in generative models (2020, methodology): “Capturing the causal structure of this process allows the different mechanisms involved in generating the data to be intervened on independently, based on the principle of Independence of Mechanisms (IM) (Janzing and Sch¨olkopf 2010; Lemeire and Janzing 2012; Peters, Janzing, and Sch¨olkopf 2017).”
• Group invariance principles for causal generative models (2017, background): “In this paper, we suggest that a group theoretic framework can unify ICM-based approaches and provide useful tools to study generative models in general.”
• Telling cause from effect in deterministic linear dynamical systems (2015, methodology): “Here, we refer to this abstract principle as the postulate of Independence of Cause and Mechanism (ICM).”
• Distinguishing Cause from Effect Using Observational Data: Methods and Benchmarks (2014, methodology): “If “complexity” and “information” are measured by Kolmogorov complexity and algorithmic information, respectively (Janzing and Schölkopf, 2010; Lemeire and Janzing, 2013), one can show that the statement “pC contains no information about pE |C” implies that the sum of the complexities of pC and pE |”
• Distinguishing Cause from Effect Using Observational Data: Methods and Benchmarks (2014, methodology): “If “complexity” and “information” are measured by Kolmogorov complexity and algorithmic information, respectively (Janzing and Schölkopf, 2010; Lemeire and Janzing, 2013), one can show that the statement “pC contains no information about pE |C” implies that the sum of the complexities of pC and pE…”
• The lesson of causal discovery algorithms for quantum correlations: causal explanations of Bell-inequality violations require fine-tuning (2012, background): “In addition, we consider only algorithms that look at statistical independences; those that use algorithmic independences are not considered [4].”
• Intrinsic disentanglement: an invariance view for deep generative models (2018, methodology): “This hypothesis relies on recent work exploiting the postulate of Independence of Cause and Mechanism stating that Nature chooses independently the properties of a cause and those of the mechanism that generate effects from the cause (Janzing and Schölkopf, 2010; Lemeire and Janzing, 2012).”
• Cause-Effect Inference by Comparing Regression Errors (2018, background): “For the purpose of foundational insights (rather than for practical purposes), [17, 18] formalize the idea via algorithmic information and postulate that knowing pC does not enable a shorter description of pE|C and vice versa.”