This paper proposes a novel method to estimate sleep stage in real-time with a non-contact device. The proposed method employs the trigonometric function regression model to estimate prospective heart rate from the partially obtained heart rate and calculates the sleep stage from the estimated heart rate. This paper conducts the subject experiment and it is revealed that the proposed method enables to estimate the sleep stage in real-time, in particular the proposed method has the equivalent estimation accuracy as the previous method that estimates the sleep stage according to the entire heart rate during sleeping.

Highly sensitive low noise electrodes, capability of fast processing of multivariate signals, low cost of hardware and wireless communication have widen the possibilities of the use of electroencephalogram (EEG) data for various applications. These applications are not only restricted to medical investigations (like epileptic seizures, monitoring anesthesia or brain functions etc.), but also for well-being of disabled patients, as well as for entertainments like playing games. Our focus in this study is BCI applications based on identification of Event Related Potential (ERP) P300. One such application is BCI speller, which is used in our experiments. BCI speller on the market use 8 probes and take 72 seconds to collect data to reliably spell a single character. The motivation of this work is to reduce the number of probes and the time needed to spell a letter. A commercial product should reliably work for every user (customer). Such a large number of probes and long time to spell a letter are necessary to ensure correct recognition. We have shown that if we identify the position of the probes appropriately for an individual, as few as two probes could give even better results. All experiments are conducted at our in-house facility, where most ot the subjects undergone no prior training.

Due to the advances in medical science, increasing health consciousness, improved quality of food, the average human life span has increased to a great extent. On the other hand, stresses of modern life, overwork and less sleep, increased usage of digital devices and internet, less exercise, are leading us to poor quality of life. Elderly people are more vulnerable to reduced life quality due to deterioration of both physical and mental health. People at any age need to maintain a minimum level of wellbeing to pursue his or her daily activities to lead a fulfilling life. Thus the need of assessing and restoring wellness is very important. Fortunately the progress of information and communication technologies provide use sensor devices and computing platform to feel, monitor and restore the wellness. In this work, a study has been done to define and determine wellness related to daily activities data obtained from various sensors and provide recommendation to the user regarding improvement of life style to achieve wellness. A small-scale experiment has been done using a simple lifelog device. The daily activities data including walking steps, sleep time, inactive period, calories burned are collected from 8 subjects. In addition food intake, eating times, cell phone usage, messaging time, time of interaction with other people and solo time are also manually collected. The correlation of physical activities (walking time, exercise time), mental activities (cell phone usage, study time, interaction with friends) and sleep patterns are studied. A simple parameter Tiredness Factor has been proposed to determine wellness and a recommendation system for improving wellness has been developed. Questionnaire from the subjects about the personal feelings of wellness has been noted and used to evaluate our proposal.

Texts posted on the web by users from diverse cultures provide a nearly endless source of data that researchers can use to study human thoughts and language patterns. However, unless care is taken to avoid it, models may be developed in one cultural setting and deployed in another, leading to unforeseen consequences. We explore the effects of using models built from a corpus of texts from multiple cultures in order to learn about each represented people group separately. To do this, we employ a topic modeling approach to quantify open-ended writing responses describing personal values and everyday behaviors in two distinct cultures. We show that some topics are more prominent in one culture compared to the other, while other topics are mentioned to similar degrees. Furthermore, our results indicate that culture influences how value-behavior relationships are exhibited. While some relationships exist in both cultural groups, in most cases we see that the observed relations are dependent on the cultural background of the data set under examination.

Numerous studies have identified differences between left-handed and right-handed people, especially in the fields of psychology and neuroscience. Using a social media setting, this paper presents a data-driven approach to explore whether a person's handedness can be identified given his or her writing, and shows handedness characteristics that can be inferred from language.

Artificially intelligent conversational agents have been demonstrated to positively impact team based learning in classrooms and hold even greater potential for impact in the now widespread Massive Open Online Courses (MOOCs) if certain challenges can be overcome. These challenges include team formation, coordination and management of group processes in teams working together while distributed both in time and space. Our work begins with an architecture for orchestrating conversational agent based support for group learning called Bazaar, which has facilitated numerous successful studies of learning in the past including some early investigations in MOOC contexts. In this paper, we briefly describe our experience in designing, developing and deploying agent supported collaborative learning activities in 3 different MOOCs in three iterations. Findings from this iterative design process provide an empirical foundation for a reusable framework for facilitating similar activities in future MOOCs.

People increasingly rely on machine learning (ML) to make intelligent decisions. However, the ML results are often difficult to interpret and the algorithms do not support interaction to solicit clarification or explanation. In this paper, we highlight an emerging research area of interpretable explanations for transfer learning in sequential tasks, in which an agent must explain how it learns a new task given prior, common knowledge. The goal is to enhance a user's ability to trust and use the system output and to enable iterative feedback for improving the system. We review prior work in probabilistic systems, sequential decision-making, interpretable explanations, transfer learning, and interactive machine learning, and identify an intersection that deserves further research focus. We believe that developing adaptive, transparent learning models will build the foundation for better human-machine systems in applications for elder care, education, and health care.

Researchers have been developing Social robots are increasingly common in schools to support algorithms to aid robots in determining task hierarchies learning goals, in workplaces to augment productivity, (Hayes and Scassellati 2014), learning tasks from humans and in homes to improve quality of life. The fulfillment of (Thomaz and Breazeal 2008), and choosing what information their objectives in these environments are strongly dependent to communicate and when to communicate it (Unhelkar on the quality of the sustained, supportive relationship and Shah 2016). Although robots have made great robots are able to construct with their human users.

Dialog between drivers and speech-based robot vehicle interfaces can be used as an instrument to find out what drivers might be concerned, confused or curious about in driving simulator studies. Eliciting ongoing conversation with drivers about topics that go beyond navigation, control of entertainment systems, or other traditional driving related tasks is important to getting drivers to engage with the activity in an open-ended fashion. In a structured improvisational Wizard of Oz study that took place in a highly immersive driving simulator, we engaged participant drivers (N=6) in an autonomous driving course where the vehicle spoke to drivers using computer-generated natural language speech. First, using microanalyses of drivers’ responses to the car’s utterances, we identify a set of topics that are expected and treated as appropriate by the participants in our study. Second, we identify a set of topics and conversational strategies that are treated as inappropriate. Third, we show that it is just these unexpected, inappropriate utterances that eventually increase users’ trust into the system, make them more at ease, and raise the system’s acceptability as a communication partner.

Methods and materials are described for employing a human-shaped robot as a lecturer in automated remote instruction. Video segments from the stimuli of a 2 (participant substrate: VR or non-VR) x 2 (robot embodiment: copresent or screen) balanced between-participants experiment are provided. In each condition, a robot delivers the content for a lecture on the nutrition of carbohydrates. The robot uses identical speech and body movement while the same set of slides plays on an adjacent computer, thereby controlling for such factors as educational content, robot appearance and robot size. The experiment employs Aldebaran’s 25-degrees-of-freedom Nao as the robot and the Oculus Rift as the immersive VR system. The lecture speech and slides were obtained with permission from a Mandarin Chinese-language online course and translated into English. The setup for different delivery modes for automated remote instruction are illustrated using a robot delivering foreign language online content. These methods support the design and evaluation of robots that perform the role of lecturer.