Deep reinforcement learning is a dynamic field that has been successfully applied in various engineering and scientific disciplines. However, like many areas of automated learning, it presents a significant challenge: understanding its models, which can hinder human trust in the decisions made by these algorithms. To address this issue, we conducted a focus group and interviews with experts in machine learning and reinforcement learning to gain insights into the perceptions and preferences surrounding interpretability techniques and tools in reinforcement learning systems. The discussion of the focus groups used an interactive dashboard to simplify the analysis of reinforcement learning environments. The focus group results showed many problems to tackle with the current approaches regarding RL agent training, like data manipulation, environment representation and reward misinterpretation. The focus group discussion also presented some key features of a visualization approach to interpretability in RL: contextual presentation of information, integration with existing pipelines, and ease of distribution.

This work describes an Augmented Reality Serious Game (ARSG) focused on facilitating the process of acquiring new skills in individuals with motor disabilities. In particular, this game aims to help them control a robotic wheelchair. A racing track was used as a game narrative, including restriction areas, static and dynamic objects, obstacles and various signs. A user study with 20 participants was conducted to compare different methods used to place virtual content on the real-world environment while the user interacts with the game and control the wheelchair in the physical space: C1 - motion tracking using cloud anchors; C2 - offline motion tracking. Results suggest condition C1 is more precise and robust, while condition C2 seems to be easier to configure.

To support the nuances of collaborative work, many researchers have been exploring the field of Augmented Reality (AR), aiming to assist in co-located or remote scenarios. Solutions using AR allow taking advantage from seamless integration of virtual objects and real-world objects, thus providing collaborators with a shared understanding or common ground environment. However, most of the research efforts, so far, have been devoted to experiment with technology and mature methods to support its design and development. Therefore, it is now time to understand where the field stands and how well can it address collaborative work with AR, to better characterize and evaluate the collaboration process. In this article, we perform an analysis of the different dimensions that should be taken into account when analysing the contributions of AR to the collaborative work effort. Then, we bring these dimensions forward into a conceptual framework and propose an extended human-centered taxonomy for the categorization of the main features of Collaborative AR. Our goal is to foster harmonization of perspectives for the field, which may help create a common ground for systematization and discussion. We hope to influence and improve how research in this field is reported by providing a structured list of the defining characteristics. Finally, some examples of the use of the taxonomy are presented to show how it can serve to gather information for characterizing AR-supported collaborative work, and illustrate its potential as the grounds to elicit further studies.

Augmented Reality (AR) has been explored with the objective to assist in scenarios of co-located or remote collaboration. To help understand how well collaborative work can be addressed with AR, it is important to foster harmonization of perspectives and create a common ground for systematization and discussion. In this vein, understand relationships among existing dimensions of collaboration, as well as identify research opportunities, is of paramount importance and thus tools that allow visually exploring information associated with Collaborative AR may be most valuable. In this paper, we present a first effort towards the creation of such an interactive visualization tool for exploration and analysis of collaborative AR research. It allows visualize data of selected papers organized according to a human-centered taxonomy on collaborative AR. In order to get insights into whether the structure was understood and if the representation was clear and efficient to use, we evaluated the proposed tool through a user study with 40 participants. Results suggest the tool has potential towards the creation of a shared understanding and identification of existing patterns, trends and opportunities within the field of collaborative AR.

Artificial Intelligent (AI) techniques, such as ma-chine learning (ML), have been making significant progress over the past decade. Many systems have been applied in sensitive tasks involving critical infrastructures which affect human well-being or health. Before deploying an AI system, it is necessary to validate its behavior and guarantee that it will continue to perform as expected when deployed in a real-world environment. For this reason, it is important to comprehend specific aspects of such systems. For example, understanding how neural networks produce final predictions remains a fundamental challenge. Existing work on interpreting neural network predictions for images via feature visualization often focuses on explaining predictions for neurons of one single convolutional layer. Not presenting a global perspective over the features learned by the model leads the user to miss the bigger picture. In this work we focus on providing a representation based on the structure of deep neural networks. It presents a visualization able to give the user a global perspective over the feature maps of a convolutional neural network (CNN) in a single image, revealing potential problems of the learning representations present in the network feature maps.

Assembly procedures are a common task in several domains of application. Augmented Reality (AR) has been considered as having great potential in assisting users while performing such tasks. However, poor interaction design and lack of studies, often results in complex and hard to use AR systems. This paper considers three different interaction methods for assembly procedures (Touch gestures in a mobile device; Mobile Device movements; 3D Controllers and See-through HMD). It also describes a controlled experiment aimed at comparing acceptance and usability between these methods in an assembly task using Lego blocks. The main conclusions are that participants were faster using the 3D controllers and Video see-through HMD. Participants also preferred the HMD condition, even though some reported light symptoms of nausea, sickness and/or disorientation, probably due to limited resolution of the HMD cameras used in the video see-through setting and some latency issues. In addition, although some research claims that manipulation of virtual objects with movements of the mobile device can be considered as natural, this condition was the least preferred by the participants.

The decision-making process and the development of decision support systems (DSS) have been enhanced by a variety of methods originated from information science, cognitive psychology and artificial intelligence over the past years. Situated visualization (SV) is a method to present data representations in context. Its main characteristic is to display data representations near the data referent. As augmented reality (AR) is becoming more mature, affordable and widespread, using it as a tool for SV becomes feasible in several situations. In addition, it may provide a positive contribution to more effective and efficient decision-making, as the users have contextual, relevant and appropriate information to endorse their choices. As new challenges and opportunities arise, it is important to understand the relevance of intertwining these fields. Based on a literature analysis, this paper addresses and discusses current areas of application, benefits, challenges and opportunities of using SV through AR to visualize data in context and to support a decision-making process and its importance in future DSS.