IBM Cloud Services Enhance Automatic Cognitive Assessment via Human-Robot Interaction

Abstract

Thanks to recent developments in artificial intelligence and social robotics, Human-Robot Interaction (HRI) can be used as a non-invasive screening tool for the assessment of cognitive decline. In this scenario, the robot manages the assessment by providing the instructions to the patient, registering his/her answers and objectively calculating the final score. This service can help to save time and reach a wider population. From the technical point of view, a challenge is to achieve a highly reliable speech and visual recognition as required for a valid scoring of performance. In this article, we evaluate a system for cognitive assessment that makes use of the IBM AI Cloud services embodied in one of the most popular platforms for social robotics: the SoftBank Pepper. Results of a pilot study with 16 human participants shows that IBM Cloud services for speech and visual recognition can improve the system performance in comparison with standard interfaces. Importantly, the improvement allows achieving a significant correlation with one of the most used paper-and-pencil tests and, therefore, the study demonstrates the validity of the robotic approach for cognitive assessment.

Psychometric Evaluation Supported by a Social Robot: Personality Factors and Technology Acceptance

“…The results and limitations of the present study will be feedback Margin requirements for the other pages Paper size this page US Letter future work where we aim to design a new trial with a larger sample to provide further evidence. In future work, we will also firther explore the use of the most advanced AI Cloud services that demonstrated to be beneficial to improve the automatic scoring of the tests [23].…”

Assessment of Cognitive skills via Human-robot Interaction and Cloud Computing

“…Solutions adopted in this work are informed by previous preliminary experiments with the SoftBank humanoid social robot Pepper [35] , it was found that the score automatically calculated using the robot software often failed to provide the correct score because of the failures of the embedded speech and object recognition interfaces [23,36] . To tackle this problem, after thorough consideration of what the market can offer, we selected the IBM AI Cloud services, named "Watson", which provide a comprehensive set of easy-to-use tools for speech recognition (speech-to-text) and production (text-to-speech) and object recognition from picture, which definitely improved the robotic system interaction and automated scoring [37] .…”

“…Utilising these features, the Watson score was calculated via a customised and flexible approach which drastically improved the system performance as reported in our preliminary study [37] . It was customized because we used the language customization service, to refine the speech-to-text model for the specific recognition requests of our cognitive test.…”

Robotics Utilization in Automatic Vision-Based Assessment Systems From Artificial Intelligence Perspective: A Systematic Review

“…The research interest in developing assistive technology to conduct real-time activities, including automated visionbased assessment, is growing. Considerable research has been performed to develop methods for the automatic detection and classification of human behaviour in various environments, such as domestic [18], education [5] [17] [19] and healthcare [20] [21]. Automated vision-based assessment is important for human-machine interactions requiring a highly accurate classifier [10].…”