Ensemble approaches to classification and regression have attracted a great deal of interest in recent years. These methods can be shown both theoretically and empirically to outperform single predictors on a wide range of tasks. One of the elements required for accurate prediction when using an ensemble is recognised to be error "diversity". However, the exact meaning of this concept is not clear from the literature, particularly for classification tasks. In this paper we first review the varied attempts to provide a formal explanation of error diversity, including several heuristic and qualitative explanations in the literature. For completeness of discussion we include not only the classification literature but also some excerpts of the rather more mature regression literature, which we believe can still provide some insights. We proceed to survey the various techniques used for creating diverse ensembles, and categorise them, forming a preliminary taxonomy of diversity creation methods. As part of this taxonomy we introduce the idea of implicit and explicit diversity creation methods, and three dimensions along which these may be applied. Finally we propose some new directions that may prove fruitful in understanding classification error diversity.

This paper presents a method for one-shot learning of dexterous grasps, and grasp generation for novel objects. A model of each grasp type is learned from a single kinesthetic demonstration, and several types are taught. These models are used to select and generate grasps for unfamiliar objects. Both the learning and generation stages use an incomplete point cloud from a depth camera -no prior model of object shape is used. The learned model is a product of experts, in which experts are of two types. The first is a contact model and is a density over the pose of a single hand link relative to the local object surface. The second is the hand configuration model and is a density over the whole hand configuration. Grasp generation for an unfamiliar object optimises the product of these two model types, generating thousands of grasp candidates in under 30 seconds. The method is robust to incomplete data at both training and testing stages. When several grasp types are considered the method selects the highest likelihood grasp across all the types. In an experiment, the training set consisted of five different grasps, and the test set of forty-five previously unseen objects. The success rate of the first choice grasp is 84.4% or 77.7% if seven views or a single view of the test object are taken, respectively. Keywords learning, dexterous grasping * Authors Kopicki, Detry and Wyatt are identified as the primary authors of this work. Kopicki is identified as the first author. † Corresponding author; e-mail: jlw@cs.bham.ac.uk 2 IJRR -(-) Fig. 1. Leftmost image: Objects used, the four objects on the left were used solely for training, the remaining forty three objects on the right were solely used as novel test objects. Rightmost image: The Boris manipulation platform on which the experiments reported were carried out.

Thanks to the efforts of the robotics and autonomous systems community, robots are becoming ever more capable. There is also an increasing demand from end-users for autonomous service robots that can operate in real environments for extended periods. In the STRANDS project we are tackling this demand head-on by integrating state-of-the-art artificial intelligence and robotics research into mobile service robots, and deploying these systems for long-term installations in security and care environments. Over four deployments, our robots have been operational for a combined duration of 104 days autonomously performing end-user defined tasks, covering 116km in the process. In this article we describe the approach we have used to enable long-term autonomous operation in everyday environments, and how our robots are able to use their long run times to improve their own performance

A long-standing goal of AI is to enable robots to plan in the face of uncertain and incomplete information, and to handle task failure intelligently. This paper shows how to achieve this. There are two central ideas. The first idea is to organize the robot's knowledge into three layers: instance knowledge at the bottom, commonsense knowledge above that, and diagnostic knowledge on top. Knowledge in a layer above can be used to modify knowledge in the layer(s) below. The second idea is that the robot should represent not just how its actions change the world, but also what it knows or believes. There are two types of knowledge effects the robot's actions can have: epistemic effects (I believe X because I saw it) and assumptions (I'll assume X to be true). By combining the knowledge layers with the models of knowledge effects, we can simultaneously solve several problems in robotics: (i) task planning and execution under uncertainty; (ii) task planning and execution in open worlds; (iii) explaining task failure; (iv) verifying those explanations. The paper describes how the ideas are implemented in a three-layer architecture on a mobile robot platform. The robot implementation was evaluated in five different experiments on object search, mapping, and room categorization.

This paper describes an architecture for robots that combines the complementary strengths of probabilistic graphical models and declarative programming to represent and reason with logic-based and probabilistic descriptions of uncertainty and domain knowledge. An action language is extended to support non-boolean fluents and nondeterministic causal laws. This action language is used to describe tightly-coupled transition diagrams at two levels of granularity, with a fine-resolution transition diagram defined as a refinement of a coarse-resolution transition diagram of the domain. The coarse-resolution system description, and a history that includes (prioritized) defaults, are translated into an Answer Set Prolog (ASP) program. For any given goal, inference in the ASP program provides a plan of abstract actions. To implement each such abstract action, the robot automatically zooms to the part of the fine-resolution transition diagram relevant to this action. A probabilistic representation of the uncertainty in sensing and actuation is then included in this zoomed fine-resolution system description, and used to construct a partially observable Markov decision process (POMDP). The policy obtained by solving the POMDP is invoked repeatedly to implement the abstract action as a sequence of concrete actions, with the corresponding observations being recorded in the coarse-resolution history and used for subsequent reasoning. The architecture is evaluated in simulation and on a mobile robot moving objects in an indoor domain, to show that it supports reasoning with violation of defaults, noisy observations and unreliable actions, in complex domains. 1 We use the terms "robot" and "agent" interchangeably in this paper. 1 arXiv:1508.03891v4 [cs.RO] 21 Sep 2018 probability reason optimally (or near optimally) about the effects of numerically quantifiable uncertainty in sensing and action. There have been many attempts to combine the benefits of these two classes of systems, including work on joint (i.e., logic-based and probabilistic) representations of state and action, and algorithms for planning and decisionmaking in such formalisms. These approaches provide significant expressive power, but they also impose a significant computational burden. More efficient (and often approximate) reasoning algorithms for such unified probabilisticlogical paradigms are being developed. However, practical robot systems that combine abstract task-level planning with probabilistic reasoning, link, rather than unify, their logic-based and probabilistic representations, primarily because roboticists often need to trade expressivity or correctness guarantees for computational speed. Information close to the sensorimotor level is often represented probabilistically to quantitatively model and reason about the uncertainty in sensing and actuation, with the robot's beliefs including statements such as "the robotics book is on the shelf with probability 0.9". At the same time, logic-based systems are used to reason with (more) abstract commonsen...

Current approaches to visual object class detection mainly focus on the recognition of abstract object categories, such as cars, motorbikes, mugs and bottles. Although these approaches have demonstrated impressive performance in terms of recognition, their restriction to abstract categories seems artificial and inadequate in the context of embodied, cognitive agents. Here, distinguishing objects according to functional aspects based on object affordances is vital for a meaningful human-machine interaction. In this paper, we propose a complete system for the detection of functional object classes, based on a representation of visually distinct hints on object affordances (affordance cues). It spans the complete cycle from tutor-driven acquisition of affordance cues, one-shot learning of corresponding object models, and detecting novel instances of functional object classes in real images

Abstract. Intrinsic images represent the underlying properties of a scene such as illumination (shading) and surface reflectance. Extracting intrinsic images is a challenging, ill-posed problem. Human performance on tasks such as shadow detection and shape-from-shading is improved by adding colour and texture to surfaces. In particular, when a surface is painted with a textured pattern, correlations between local mean luminance and local luminance amplitude promote the interpretation of luminance variations as illumination changes. Based on this finding, we propose a novel feature, local luminance amplitude, to separate illumination and reflectance, and a framework to integrate this cue with hue and texture to extract intrinsic images. The algorithm uses steerable filters to separate images into frequency and orientation components and constructs shading and reflectance images from weighted combinations of these components. Weights are determined by correlations between corresponding variations in local luminance, local amplitude, colour and texture. The intrinsic images are further refined by ensuring the consistency of local texture elements. We test this method on surfaces photographed under different lighting conditions. The effectiveness of the algorithm is demonstrated by the correlation between our intrinsic images and ground truth shading and reflectance data. Luminance amplitude was found to be a useful cue. Results are also presented for natural images.

Deployment of robots in practical domains poses key knowledge representation and reasoning challenges. Robots need to represent and reason with incomplete domain knowledge, acquiring and using sensor inputs based on need and availability. This paper presents an architecture that exploits the complementary strengths of declarative programming and probabilistic graphical models as a step toward addressing these challenges. Answer Set Prolog (ASP), a declarative language, is used to represent, and perform inference with, incomplete domain knowledge, including default information that holds in all but a few exceptional situations. A hierarchy of partially observable Markov decision processes (POMDPs) probabilistically models the uncertainty in sensor input processing and navigation. Nonmonotonic logical inference in ASP is used to generate a multinomial prior for probabilistic state estimation with the hierarchy of POMDPs. It is also used with historical data to construct a beta (meta) density model of priors for metareasoning and early termination of trials when appropriate. Robots equipped with this architecture automatically tailor sensor input processing and navigation to tasks at hand, revising existing knowledge using information extracted from sensor inputs. The architecture is empirically evaluated in simulation and on a mobile robot visually localizing objects in indoor domains.