Causality

By Giri Narasimhan. 2019

In philosophy intuition is used in reasoning as a test-bed for the conclusions of philosophical arguments. Logic, rhetoric and intuition are the main conceptual tools in philosophical reasoning. Intuition often acts as a sort of empirical verification of the acceptability of a particular thesis. Rather like a sort of empirical test or an experimental control, to use an analogy with what happens in natural science. The basis for this method is that intuition is generalisable, or in other words, broadly speaking, it can be shared at a universal level. Moreover, intuition must have foundational validity, a primary capacity for justification that is greater than any other alternative information. It should be greater than the reference to data from the cultural and religious tradition, for example, or the recourse to the theses of classical authors. Likewise it should be able to withstand the hypotheses and empirical confirmations of scientific and technical knowledge. Experimental philosophy appears to question intuition’s alleged foundational and universal nature. Intuition is a psychological phenomenon linked to what is conventionally known, according to some authors (Stanovich 1999; see Chap. 9 of Viale 2012), but not to others (Gigerenzer 2007), as System 1 of mind. Contrary to System 2, which is rational and explicit, this system is implicit and highly contextdependent. It is permeable to the influences of emotional variables derived from the cultural and environmental context. Seen in this way, it would seem difficult to affirm the thesis of the universality of human intuition. The underlying hypothesis derived from the findings of cognitive science argues the contrary: namely that intuition is local and contingent, changing in relation not only to cultural context but also to individual psychological variables, like personality traits or emotional and affective contingencies. Experimental philosophy has explored the universality

Subfields: Causal Inference & Discovery

How researchers see this paper

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

In Causal Inference & Discovery

Citations in the corpus

Shaking the foundations: delusions in sequence models for interaction and control (2021, background): “When Θ is observed, then conditioning or intervening on the self-generated action leads to the same prediction over 𝑂 . evidence about the causal future, but erroneously also about the causal past (Figure 2a).”
Explainable Artificial Intelligence (XAI): Concepts, Taxonomies, Opportunities and Challenges toward Responsible AI (2019, neutral): “The inference of causal relationships from observational data is a field that has been broadly studied over time [161].”
Causal Effect Inference with Deep Latent-Variable Models (2017, positive): “where equality (i) is by the rules of do-calculus applied to the causal graph in Figure 1 [40].”
Improving the accuracy of medical diagnosis with causal machine learning (2020, methodology): “Despite the wealth of literature placing causal reasoning at the centre of diagnosis, to the best of our knowledge there are no existing approaches to model-based diagnosis that employ modern causal analysis techniques 49,50 .”
Algorithmic recourse under imperfect causal knowledge: a probabilistic approach (2020, methodology): “We denote the counterfactual variable by X ( do ( X I = θ )) | x F , whose distribution can be computed in three steps [27]: 1.”
Time Series Deconfounder: Estimating Treatment Effects over Time in the Presence of Hidden Confounders (2019, methodology): “However, these works are based on Pearl’s causal framework (Pearl, 2009) and use structural equation models, while our method deconfounds the estimation of treatment effects in the potential outcomes framework (Neyman, 1923; Rubin, 1978; Robins & Hernán, 2008).”
Generalization in anti-causal learning (2018, background): “We assume that the causal mechanism P ( Y | X ) is invariant with respect to changes in the distribution of the cause P ( X ) (Pearl, 2009; Peters et al., 2017), i.e., P ( Y | X ) is a physical mechanism that applies for any input distribution P ( X ) on X . the Y This is not true for anti-causal…”
Deconfounding Reinforcement Learning in Observational Settings (2018, methodology): “We recommend Pearl’s excellent monograph for further reading (Pearl, 2009).”
Counterfactual Normalization: Proactively Addressing Dataset Shift and Improving Reliability Using Causal Mechanisms (2018, background): “In addition to the assumptions for estimating potential outcomes, computing counterfactual queries requires functional or structural knowledge (Pearl, 2009).”
Evaluating Reinforcement Learning Algorithms in Observational Health Settings (2018, methodology): “While formal statistical methods have been developed to assess the quality of new policies based on observational data alone [Thomas and Brunskill, 2016, Precup et al., 2000, Pearl, 2009, Imbens and Rubin, 2015], these methods rely on strong assumptions and are limited by statistical properties.”
Fast Conditional Independence Test for Vector Variables with Large Sample Sizes (2018, background): “Many extant causal discovery algorithms use conditional independence tests to infer the underlying causal structure over a set of variables from observational data [Pearl, 2000, Spirtes et al., 2000, Hoyer et al., 2009].”
Causal Generative Neural Networks (2017, positive): “The resulting joint distribution noted P do ( X i = v i ) ( X ) , called interventional distribution [Pearl, 2009], can be computed from the FCM by discarding all causal inﬂuences on X i and clamping its value to v i .”
Explaining Classifiers with Causal Concept Effect (CaCE) (2019, methodology): “In classic tasks of causal inference from observational data, the usual methods for estimating causal effects rely either on adjusting for all confounders, known as backdoor adjutmsnet [12], or on identifying natural experiments and exploiting them, as is the case for instrumental variable analysis.”
The Selective Labels Problem: Evaluating Algorithmic Predictions in the Presence of Unobservables (2017, background): “is problem has been studied extensively in the context of causal inference where the goal is to determine the eect of some treatment [2, 30, 32, 37, 38].”
Deep and Hierarchical Implicit Models (2017, background): “They also underlie structural equation models in causality (Pearl, 2000).”
Automated versus Do-It-Yourself Methods for Causal Inference: Lessons Learned from a Data Analysis Competition (2017, background): “In the language of causal graphs (Pearl, 2009; Pearl et al., 2009), we can say that in our data generating process, the set of observed covariates X form an admissible set for a back-door path from the outcome Y to the treatment Z.”