As text classifiers become increasingly used in real-time applications, it is critical to consider not only their accuracy but also their robustness to changes in the data distribution. In this paper, we consider the case where there is a confounding variable Z that influences both the text features X and the class variable Y . For example, a classifier trained to predict the health status of a...

While conventional approaches to causal inference are mainly based on conditional (in)dependences, recent methods also account for the shape of (conditional) distributions. The idea is that the causal hypothesis “X causes Y ” imposes that the marginal distribution PX and the conditional distribution PY |X represent independent mechanisms of nature. Recently it has been postulated that the short...

Must causal models distinguish default from deviant events? Much recent work on actual causation is conducted within the structural equations framework (Spirtes et al 1993, Pearl 2000), via the notion of a causal model. In standard causal models one sets up a system of variables, allots values to these variables, and connects these variables via structural equations. Menzies (2004, 2007), Hitch...

Wand for many valuable discussions on these topics. All errors are my responsibility.

The verbs cause, enable, and prevent express beliefs about the way the world works. We offer a theory of their meaning in terms of the structure of those beliefs expressed using qualitative properties of causal models, a graphical framework for representing causal structure. We propose that these verbs refer to a causal model relevant to a discourse and that "A causes B" expresses the belief th...

Reasoning by jurors concerning whether an accused person should be convicted of committing a crime is a kind of casual inference. Jurors need to decide whether the evidence in the case was caused by the accused’s criminal action or by some other cause. This paper compares two computational models of casual inference: explanatory coherence and Bayesian networks. Both models can be applied to leg...

In an increasingly common class of studies, the goal is to evaluate causal effects of treatments that are only partially controlled by the investigator. In such studies there are two conflicting features: (1) a model on the full design and data can identify the causal effects of interest, but the model’s use in extreme regions of the data (e.g., where the outcome of interest is rare) can be sen...

Tractability analysis in terms of the causal graphs of planning problems has emerged as an important area of research in recent years, leading to new methods for the derivation of domain-independent heuristics (Katz and Domshlak 2010). Here we continue this work, extending our knowledge of the frontier between tractable and NP-complete fragments. We close some gaps left in previous work, and in...

Hausman & Woodward present an argument for the Causal Markov Condition (CMC) on the basis of a principle they dub ‘modularity’ ([1999, 2004]). I show that the conclusion of their argument is not in fact the CMC but a substantially weaker proposition. In addition, I show that their argument is invalid and trace this invalidity to two features of modularity, namely, that it is stated in terms of ...