Appearance is not everything - Preferred feature combinations for care robots

With the advancements in AI, agents (i.e., smart products, robots, software agents) are increasingly capable of working closely together with humans in a variety of ways while benefiting from each other. These human-agent collaborations have gained growing attention in the HCI community; however, the field lacks clear guidelines on how to design the agents’ behaviors in collaborations. In this paper, the qualities that are relevant for designers to create robust and pleasant human-agent collaborations were investigated. Bratman's Shared Cooperative Activity framework was used to identify the core characteristics of collaborations and survey the most important issues in the design of human-agent collaborations, namely code-of-conduct, task delegation, autonomy and control, intelligibility, common ground, offering help and requesting help. The aim of this work is to add structure to this growing and important facet of HCI research and operationalize the concept of human-agent collaboration with concrete design considerations.

/pdf/designing-human-agent-collaborations-commitment-2k01h3l3.pdf

Designing Human-Agent Collaborations: Commitment, responsiveness, and support

Artificial Intelligence in Healthcare: Review, Ethics, Trust Challenges & Future Research Directions

This paper proposes a three-dimensional (3D) point-of-intention (POI) determination method using multimodal fusion between hand pointing and eye gaze for a 3D virtual display. In the method, the finger joint forms of the pointing hand sensed by a Leap Motion sensor are first detected as pointing intention candidates. Subsequently, differences with neighboring frames, which should be during hand pointing period, are checked by AND logic with the hand-pointing intention candidates. A crossing point between the eye gaze and hand pointing lines is finally decided by the closest distance concept. In order to evaluate the performance of the proposed method, experiments with ten participants, in which they looked at and pointed at nine test points for approximately five second each, were performed. The experimental results show the proposed method measures 3D POIs at 75 cm, 85 cm, and 95 cm with average distance errors of 4.67%, 5.38%, and 5.71%, respectively.

3D Point-of-Intention Determination Using a Multimodal Fusion of Hand Pointing and Eye Gaze for a 3D Display

Service robots with advanced intelligence capabilities can potentially transform servicescapes. However, limited attention has been given to how consumers experiencing vulnerabilities, particularly those with disabilities, envisage the characteristics of robots’ prospective integration into emotionally intense servicescapes, such as long-term care (LTC). We take an interdisciplinary approach conducting three exploratory studies with consumers with disabilities involving Community Philosophy, LEGO ® Serious Play ® , and Design Thinking methods. Addressing a lack of consumer-centric research, we offer a three-fold contribution by 1) developing a conceptualization of consumer-conceived value of robots in LTC, which are envisaged as a supporting resource offering consumers opportunities to realize value; 2) empirically evidencing pathogenic vulnerabilities as a potential value-destruction factor to underscore the importance of integrating service robots research with a service inclusion paradigm; and 3) providing a theoretical extension and clarification of prior characterizations of robots’ empathetic and emotion-related AI capabilities. Consumers with disabilities conceive robots able to stimulate and regulate emotions by mimicking cognitive and behavioral empathy, but unable to express affective and moral empathy, which is central to care experience. While providing support for care practices, for the foreseeable future, service robots will not, in themselves, actualize the experience of “being cared for.”

https://doi.org/10.1177/10946705221110849

Service Robots in Long-Term Care: A Consumer-Centric View

In recent years, with the current advancements in Robotics and Artificial Intelligence (AI), robots have the potential to support the field of healthcare. Robotic systems are often introduced in the care of the elderly, children, and persons with disabilities, in hospitals, in rehabilitation and walking assistance, and other healthcare situations. In this survey paper, the recent advances in robotic technology applied in the healthcare domain are discussed. The paper provides detailed information about state-of-the-art research in care, hospital, assistive, rehabilitation, and walking assisting robots. The paper also discusses the open challenges healthcare robots face to be integrated into our society.

A Survey of Robots in Healthcare

In domains such as Human-Robot Collaboration artificial agents must be able to support mutual adaptation and learning. Towards this direction, we use a discrete Soft Actor-Critic agent on a real-time collaborative game with humans. We examine how different allocations of on-line and off-line gradient updates impact the game performance and the total training time. Our results suggest that early allocation of a high number of off-line g/u can accelerate learning while shortening training duration.

https://dl.acm.org/doi/pdf/10.1145/3453892.3454004

Accelerating Human-Agent Collaborative Reinforcement Learning

The use of Reinforcement Learning (RL) approaches for dialogue policy optimization has been the new trend for dialogue management systems. Several methods have been proposed, which are trained on dialogue data to provide optimal system response. However, most of these approaches exhibit performance degradation in the presence of noise, poor scalability to other domains, as well as performance instabilities. To overcome these problems, we propose a novel approach based on the incremental, sample-efficient Least-Squares Policy Iteration (LSPI) algorithm, which is trained on compact, fixed-size dialogue state encodings, obtained from deep Variational Denoising Autoencoders (VDAE). The proposed scheme exhibits stable and noise-robust performance, which significantly outperforms the current state-of-the-art, even in mismatched noise environments.

Variational Denoising Autoencoders and Least-Squares Policy Iteration for Statistical Dialogue Managers

Statistical Dialogue Systems (SDS) have proved their humongous potential over the past few years. However, the lack of efficient and robust representations of the belief state (BS) space refrains them from revealing their full potential. There is a great need for automatic BS representations, which will replace the old hand-crafted, variable-length ones. To tackle those problems, we introduce a novel use of Autoencoders (AEs). Our goal is to obtain a low-dimensional, fixed-length, and compact, yet robust representation of the BS space. We investigate the use of dense AE, Denoising AE (DAE) and Variational Denoising AE (VDAE), which we combine with GP-SARSA to learn dialogue policies in the PyDial toolkit. In this framework, the BS is normally represented in a relatively compact, but still redundant summary space which is obtained through a heuristic mapping of the original master space. We show that all the proposed AE-based representations consistently outperform the summary BS representation. Especially, as the Semantic Error Rate (SER) increases, the DAE/VDAE-based representations obtain state-of-the-art and sample efficient performance.

/pdf/robust-belief-state-space-representation-for-statistical-2vrfnnp1is.pdf

Robust Belief State Space Representation for Statistical Dialogue Managers Using Deep Autoencoders

Safe and efficient Human-Robot Collaboration (HRC) requires recognizing human collaborators intention Hand pose kinematics early on an ongoing movement can provide information for predicting future human actions Accurate state-of-the-art methods used for human hand pose estimation are either marker-based or make use of multiple cameras set around the workspace These approaches introduce inconvenience to the user, necessitate calibration and are bounded to the specific set-up and workspace On the other hand, using a single RGB-D camera would be less obtrusive for the user and less cumbersome to install In this work, we use OpenPose to extract 2D keypoints from the RGB raw image and we combine them with the depth information acquired from the RGB-D camera to obtain 3D hand poses We evaluate the accuracy and discrimination ability of our method in ten different static poses

Fotios Lygerakis

Papers

A Survey of Robots in Healthcare

Accelerating Human-Agent Collaborative Reinforcement Learning

Variational Denoising Autoencoders and Least-Squares Policy Iteration for Statistical Dialogue Managers

Robust Belief State Space Representation for Statistical Dialogue Managers Using Deep Autoencoders

Evaluation of 3D markerless pose estimation accuracy using openpose and depth information from a single RGB-D camera