21 research outputs found
Réseaux de Neurones Génératifs pour la Découverte de Mécanismes Causaux: Algorithmes et Applications
No full textCausal discovery is of utmost importance for agents who must plan, reason anddecide based on observations; where mistaking correlation with causation mightlead to unwanted consequences. The goldstandard to discover causal relations is to perform experiments.However, experiments are in many cases expensive, unethical, or impossible torealize. In these situations, there is a need for observational causaldiscovery, that is, the estimation of causal relations from observations alone. Causal discovery in the observational data setting traditionally involves making significant assumptions on the data and on the underlying causal model.This thesis aims to alleviate some of the assumptions made on the causal models by exploiting the modularity and expressivenessof neural networks for causal discovery, leveraging both conditionalindependences and simplicity of the causal mechanisms through two algorithms.Extensive experimentson both simulated and real-world data and a throughout theoretical anaylsisprove the good performance and the soundness of the proposedapproaches.La découverte de relations causales est primordiale pour la planification,le raisonnement et la decision basée sur des données d'observations; confondrecorrelation et causalité ici peut mener à des conséquences indésirables. Laréférence pour la découverte de relations causales est d'effectuer desexpériences contrôlées. Mais dans la majorité des cas, ces expériences sontcoûteuses, immorales ou même impossibles à réaliser. Dans ces cas, il estnécessaire d'effectuer la découverte causale seulement sur des donnéesd'observations.Dans ce contexte de causalité observationnelle, retrouver des relations causalesintroduit traditionellement des hypothèses considérables sur les données et surle modèle causal sous-jacent.Cette thèse vise à relaxer certaines de ces hypothèses en exploitant lamodularité et l'expressivité des réseaux de neurones pour la causalité, enexploitant à la fois et indépendences conditionnelles et la simplicité desméchanismes causaux, à travers deux algorithmes. Des expériences extensives surdes données simulées et sur des données réelles ainsi qu'une analyse théoriqueapprofondie prouvent la cohérence et bonne performance des approches proposées
Learning Bivariate Functional Causal Models
No full textInternational audienceFinding the causal direction in the cause-effect pair problem has been addressed in the literature by comparing two alternative generative models X → Y and Y → X. In this chapter, we first define what is meant by generative modeling and what are the main assumptions usually invoked in the literature in this bivariate setting. Then we present the theoretical identifiability problem that arises when considering causal graph with only two variables. It will lead us to present the general ideas used in the literature to perform a model selection based on the evaluation of a complexity/fit trade-off. Three main families of methods can be identified: methods making restrictive assumptions on the class of admissible causal mechanism, methods computing a smooth trade-off between fit and complexity and methods exploiting independence between cause and mechanism
Causal Discovery Toolbox: Uncovering causal relationships in Python
No full textInternational audienceThis paper presents a new open source Python framework for causal discovery from observational data and domain background knowledge, aimed at causal graph and causal mechanism modeling. The Cdt package implements an end-to-end approach, recovering the direct dependencies (the skeleton of the causal graph) and the causal relationships between variables. It includes algorithms from the 'Bnlearn' (Scutari, 2018) and 'Pcalg' (Kalisch et al., 2018) packages, together with algorithms for pairwise causal discovery such as ANM (Hoyer et al., 2009). Cdt is available under the MIT License at https://github.com/FenTechSolutions/CausalDiscoveryToolbox
Causal Discovery Toolbox: Uncovering causal relationships in Python
No full textInternational audienceThis paper presents a new open source Python framework for causal discovery from observational data and domain background knowledge, aimed at causal graph and causal mechanism modeling. The Cdt package implements an end-to-end approach, recovering the direct dependencies (the skeleton of the causal graph) and the causal relationships between variables. It includes algorithms from the 'Bnlearn' (Scutari, 2018) and 'Pcalg' (Kalisch et al., 2018) packages, together with algorithms for pairwise causal discovery such as ANM (Hoyer et al., 2009). Cdt is available under the MIT License at https://github.com/FenTechSolutions/CausalDiscoveryToolbox
Evaluation Methods of Cause-Effect Pairs
No full textInternational audienceThis chapter addresses the problem of benchmarking causal models or validating particular putative causal relationships, in the limited setting of cause-effect pairs, when empirical “observational” data are available. We do not address experimental validations e.g. via randomized controlled trials. Our goal is to compare methods, which provide a score C(X, Y ), called causation coefficient, rating a pair of variable (X, Y ) for being in a potential causal relationship X → Y . Causation coefficients may be used for various purposes, including to prioritize experiments, which may be costly or risky, or guiding decision makers in domains in which experiments are infeasible or unethical. We provide a methodology to evaluate their reliability. We take three points of views: (1) that of algorithm developers who must justify the soundness of their method, particularly with respect to identifiability and consistency, (2) that of practitioners who seek to understand on what basis algorithms make their decisions and evaluate their statistical significance, and (3) that of benchmark organizers who desire to make fair evaluations to compare methods. We adopt the framework of pattern recognition in which pairs of variable (X, Y ) and their ground truth causal graph are drawn i.i.d. from a “mother distribution”. This leads us to define new notions of probabilistic identifiability, Bayes optimal causation coefficients, and multi-part statistical tests. These new notions are evaluated on the data of the first cause-effect pair challenge. We also compile a list of resources, including datasets of real or synthetic pairs, and data generative models
Discriminant Learning Machines
No full textInternational audienceThe cause-effect pair challenge has, for the first time, formulated the cause-effect problem as a learning problem in which a causation coefficient is trained from data. This can be thought of as a kind of meta learning. This chapter will present an overview of the contributions in this domain and state the advantages and limitations of the method as well as recent theoretical results (learning theory/mother distribution). This chapter will point to code from the winners of the cause-effect pair challenge
Conditions objectives de travail et ressenti des individus : le rôle du management
No full textSynthèse n. 14 de "La Fabrique de l'Industrie"Who are the French workers? In what conditions do they work and how do they feel at work? Our exploitation of the latest Dares Working Conditions Survey (2013) answers these questions. Several profiles of individuals are thus defined by considering on the one hand their objective working conditions (working hours, remuneration, nuisances ...) and, on the other hand, their feelings about work (well-being, tensions, pride , feeling of being more or less well paid ...). This approach reveals, among other things, that one in five employees are exposed to difficult working conditions. We also note that management plays a key role in the employee's quality of life at work, which deteriorates very strongly when tensions arise with his or her management. We also note that job satisfaction obeys complex rules - that the company does not control all - and that the autonomy of the employee has a positive influence on his quality of life at work, but up to a certain threshold - after which it becomes an additional source of tension
Learning Functional Causal Models with Generative Neural Networks
No full textInternational audienceWe introduce a new approach to functional causal modeling from observational data, called Causal Generative Neural Networks (CGNN). CGNN leverages the power of neural networks to learn a generative model of the joint distribution of the observed variables, by minimizing the Maximum Mean Discrepancy between generated and observed data. An approximate learning criterion is proposed to scale the computational cost of the approach to linear complexity in the number of observations.The performance of CGNN is studied throughout three experiments.Firstly, CGNN is applied to cause-effect inference, where the task is to identify the best causal hypothesis out of and . Secondly, CGNN is applied to the problem of identifying v-structures and conditional independences. Thirdly, CGNN is applied to multivariate functional causal modeling: given a skeleton describing the direct dependences in a set of random variables , CGNN orients the edges in the skeleton to uncover the directed acyclic causal graph describing the causal structure of the random variables.On all three tasks, CGNN is extensively assessed on both artificial and real-world data, comparing favorably to the state-of-the-art. Finally, CGNN is extended to handle the case of confounders, where latent variables are involved in the overall causal model
SAM: Structural Agnostic Model, Causal Discovery and Penalized Adversarial Learning
No full textWe present the Structural Agnostic Model (SAM), a framework to estimate end-to-end non-acyclic causal graphs from observational data. In a nutshell, SAM implements an adversarial game in which a separate model generates each variable, given real values from all others. In tandem, a discriminator attempts to distinguish between the joint distributions of real and generated samples. Finally, a sparsity penalty forces each generator to consider only a small subset of the variables, yielding a sparse causal graph. SAM scales easily to hundreds variables. Our experiments show the state-of-the-art performance of SAM on discovering causal structures and modeling interventions, in both acyclic and non-acyclic graphs
: Comprendre la qualité de vie au travail
No full textNational audienceWho are the French workers? In what conditions do they work and how do they feel at work? Our exploitation of the latest Dares Working Conditions Survey (2013) answers these questions. Several profiles of individuals are thus defined by considering on the one hand their objective working conditions (working hours, remuneration, nuisances ...) and, on the other hand, their feelings about work (well-being, tensions, pride , feeling of being more or less well paid ...). This approach reveals, among other things, that one in five employees are exposed to difficult working conditions. We also note that management plays a key role in the employee's quality of life at work, which deteriorates very strongly when tensions arise with his or her management. We also note that job satisfaction obeys complex rules - that the company does not control all - and that the autonomy of the employee has a positive influence on his quality of life at work, but up to a certain threshold - after which it becomes an additional source of tension.Qui sont les travailleurs français ? Dans quelles conditions travaillent-ils et comment se sentent-ils au travail ? Notre exploitation de la dernière enquête Conditions de travail de la Dares (2013) permet de répondre à ces questions. Plusieurs profils d’individus sont ainsi définis en considérant d’une part leurs conditions de travail objectives (horaires, rémunération, exposition aux nuisances…) et, d’autre part, leur ressenti par rapport au travail (bien-être, tensions, fierté, sentiment d’être plus ou moins bien payé…).Cette approche révèle entre autres qu’un actif occupé sur cinq est exposé à des conditions de travail difficiles. On note également que le management tient un rôle essentiel dans la qualité de vie au travail du salarié, qui se dégrade très fortement lorsque surviennent des tensions avec sa hiérarchie. On remarque aussi que la satisfaction au travail obéit à des règles complexes–que l’entreprise ne maîtrise pas toutes – et que l’autonomie du salarié exerce une influence positive sur sa qualité de vie au travail, mais jusqu'à un certain seuil - après lequel elle se transforme en une source additionnelle de tensions
