Affordances encode relationships between actions, objects, and effects. They play an important role on basic cognitive capabilities such as prediction and planning. We address the problem of learning affordances through the interaction of a robot with the environment, a key step to understand the world properties and develop social skills. We present a general model for learning object affordances using Bayesian networks integrated within a general developmental architecture for social robots. Since learning is based on a probabilistic model, the approach is able to deal with uncertainty, redundancy, and irrelevant information. We demonstrate successful learning in the real world by having an humanoid robot interacting with objects. We illustrate the benefits of the acquired knowledge in imitation games.

We show how to significantly speed up MCMC sampling of DAG structures by using a powerful non-local proposal based on Koivisto’s dynamic programming (DP) algorithm (11; 10), which computes the exact marginal posterior edge probabilities by analytically summing over orders. Furthermore, we show how sampling in DAG space can avoid subtle biases that are introduced by approaches that work only with orders, such as Koivisto’s DP algorithm and MCMC order samplers (6; 5). 1

Abstract — Affordances represent the behavior of objects in terms of the robot’s motor and perceptual skills. This type of knowledge plays a crucial role in developmental robotic systems, since it is at the core of many higher level skills such as imitation. In this paper, we propose a general affordance model based on Bayesian networks linking actions, object features and action effects. The network is learnt by the robot through interaction with the surrounding objects. The resulting probabilistic model is able to deal with uncertainty, redundancy and irrelevant information. We evaluate the approach using a real humanoid robot that interacts with objects. I.

We outline a representation for discrete multivariate distributions in terms of interventional potential functions that are globally normalized. This representation can be used to model the effects of interventions, and the independence properties encoded in this model can be represented as a directed graph that allows cycles. In addition to discussing inference and sampling with this representation, we give an exponential family parametrization that allows parameter estimation to be stated as a convex optimization problem; we also give a convex relaxation of the task of simultaneous parameter and structure learning using group ℓ1regularization. The model is evaluated on simulated data and intracellular flow cytometry data. 1

Given observations about a collection of nodes, the goal of graph reconstruction is to predict a set of edges that connect the nodes in a realistic fashion. A number of similar formulations of the problem have been introduced across research areas, notably social sciences, epidemiology and biology. What is common across domains is the understanding that interesting properties of the system are known to depend on the graph structure. As a example, consider the cell signaling network previously studied in [4, 2]. Gene expression modulations are a by-product of a living network of activity. In this application, the expression levels of 11 signaling molecules have been measured using a technology called flow cytometry, and the goal is to recovery the causal influences of one molecule on another in this network. An important aspect of this data is the inclusion of interventions, whereby external factors are used to perturb the expressions of individual molecules, as these help to identify the directionality of influence between two molecules. The authors use Bayesian networks to infer the causal relations. In this framework, the expression levels of signaling molecules in a cell are modeled as discrete random variables whose dependency structure is assumed to form a directed acyclic graph. Then, using this well-established probabilistic

Causal learning methods are often evaluated in terms of their ability to discover a true underlying directed acyclic graph (DAG) structure. However, in general the true structure is unknown and may not be a DAG structure. We therefore consider evaluating causal learning methods in terms of predicting the effects of interventions on unseen test data. Given this task, we show that there exist a variety of approaches to modeling causality, generalizing DAG-based methods. Our experiments on synthetic and biological data indicate that some non-DAG models perform as well or better than DAG-based methods at causal prediction tasks.

Causal learning methods are often evaluated in terms of their ability to discover a true underlying directed acyclic graph (DAG) structure. However, in general the true structure is unknown and may not be a DAG structure. We therefore consider evaluating causal learning methods in terms of predicting the effects of interventions on unseen test data. Given this task, we show that there exist a variety of approaches to modeling causality, generalizing DAG-based methods. Our experiments on synthetic and biological data indicate that some non-DAG models perform as well or better than DAG-based methods at causal prediction tasks.

Causal learning methods are often evaluated in terms of their ability to discover a true underlying DAG structure. However, in general the true structure is unknown and may not be a DAG. We therefore consider evaluating causal learning methods in terms of predicting the effects of interventions on unseen test data. Given this task, we show that there exist a variety of approaches to modeling causality, generalizing DAG-based methods. Our experiments on synthetic and biological data indicate that some non-DAG models perform as well or better than DAG-based methods at causal prediction tasks.

Abstract. To learn about causal relations between variables just by observing samples from them, particular assumptions must be made about those variables ’ distributions. This article gives a practical description of how such a learning task can be undertaken based on different possible assumptions. Two categories of assumptions lead to different methods, constraint-based and Bayesian learning, and in each case we review both the basic ideas and some recent extensions and alternatives to them. 1

Machine learning techniques are becoming increasingly useful in the study of complex biological phenomena. As more is understood about biological regulation, and additional experimental methods are developed, it is also becoming feasible to attempt to “reverse engineer ” the pathways regulating biological systems using graphical models, with an emphasis on Bayesian networks. Herein we describe the background of Bayesian networks being applied to biological networks, apply Bayesian network inference to five-node differential equation models of two cellular networks, and discuss issues moving forward with the application of Bayesian networks to modeling phosphoproteomic data. 1