Studies of data collection in the workplace or educational environments have highlighted concerns around privacy, autonomy, and negative performance impacts of unmitigated data surveillance. However, key differences in the perspectives of that surveillance by workers and students have shown that while workers are reticent to the increasing workforce surveillance, students tend to not be concerned due to the perceived or assumed beneficence of educational institutions. In the following study, we look at a population at the intersection of workers and students: surgical residents. Like many health-professional trainees, this population is training on-the-job and there is an increasing number and forms of data collection systems being deployed as a part of that training in the name of performance improvement, feedback, and assessment. We conducted in-depth interviews with 14 surgical residents enrolled in 7 general surgery programs in the United States to investigate their privacy perceptions to determine how they may differ from those that are solely workers or students. Our findings reveal that they are unaware of how their data is shared and are concerned about data misrepresentation, fear of misuse affecting job security, and potential discrimination, but at the same time are trusting their educators to have access to their data to provide better training outcomes. Through this empirical evidence, we highlight how medical trainees have competing needs for data privacy and use compared to those who are solely workers or students. To this end, we provide initial recommendations to address privacy concerns, correct misunderstandings about data use, and promote the adoption of privacy-aware practices in health professional training.

Paramedics are often involved in varied and complex, emotionally provoking emergency calls which can result in difficulty controlling their affective experience. As a result, their internal physiological state may “leak” out through their external visual and auditory behaviors which can affect patient care. This research aims to identify how this ‘leakage’ may be perceived by observers, what commonalities exist in how the affective dimensions seem to be expressed through the body and the relationship between these dimensions and trainees' electrodermal activity (EDA). We conducted a preliminary study with a small set of knowledgeable observers to continuously rate trainees' valence, arousal and dominance from behavioral data in thin slices of simulation videos. We analyzed the relationship between trainees' EDA and observers' independent ratings. Our findings show a significant agreement on and correlation between the observers' ratings for all dimensions and preliminary modeling indicates a significant relationship.

There is a long-standing interest in CSCW on distributed instruction - both in how it differs from collocated instruction as well as the design of tools to reduce any deficiencies. In this study, we leveraged the unique environment of laparoscopic surgery to compare the efficacy and mechanism of instruction in a collocated and distributed condition. By implementing the same instructional technology in both conditions, we are able to evaluate the effect of distance on instruction without the confounding variable of medium of instruction. Surprisingly, our findings revealed trainees perceived a higher perceived quality of instruction in the distributed condition. Further investigation suggests that in a distributed learning environment, trainees change their behavior to attend more to the provided instructions resulting in this higher perceived quality of instruction. Finally, we discuss our findings with regards to media compensation theory, and we provide both social and technical insights on how to better support a distributed instructional process.

Elsevier’s Scopus, the largest abstract and citation database of peer-reviewed literature. Search and access research from the science, technology, medicine, social sciences and arts and humanities fields.

Multi-user input over a shared display has been shown to support group process and improve performance. However, current gesturing systems for instructional collaborative tasks limit the input to experts and overlook the needs of novices in making references on a shared display. In this paper, we investigate the effects of a single-user gesturing tool on the communication between trainer and trainees in a laparoscopic surgical training. By comparing the communication structure and content between the trainings with and without the gesturing tool, we show that the communication becomes more imbalanced and the trainees become less active when using the single-user gesturing tool. Our findings highlight the needs to grant all parties the same level of access to a shared display and suggest further directions in designing a shared display for instructional collaborative tasks.

Researchers have shown both performance drawbacks and benefits of using telepointers or similar display overlay-technologies in remote instruction; however, there is not a clear understanding of why there are these performance effects. This poses a challenge in knowing how and when to successfully use or design telepointing technologies in remote instruction. A better understanding is needed with the rise of remote workers in a wide array of industries from oil rig repair to surgery, and the proliferation of heads-up displays or telecommunications devices to support these future work practices. In this study, we explore how the information conveyed through a telepointer is taken up and acted upon by surgical trainees in a laparoscopic surgical telementoring setting. We collected audio and video data of 12 surgical trainees who performed standard laparoscopic surgical tasks on a physical model under the guidance of a surgical trainer. We investigated both action and talk to determine how the telepointer-based information was used. Our findings reveal three main challenges in using the instructional information conveyed through the telepointer including the trainees' tendency of attending to the telepointer instruction as the primary source of information. We argue that the found challenges are socio-technical in nature and require a redesign of the mentoring context as well as the technological tools.

Predicting the exciting portions of a video is a widely relevant problem because of applications such as video summarization, searching for similar videos, and recommending videos to users. Researchers have proposed the use of physiological indices such as pupillary dilation as a measure of emotional arousal. The key problem with using the pupil to measure emotional arousal is accounting for pupillary response to brightness changes. We propose a linear model of pupillary light reflex to predict the pupil diameter of a viewer based only on incident light intensity. The residual between the measured pupillary diameter and the model prediction is attributed to the emotional arousal corresponding to that scene. We evaluate the effectiveness of this method of factoring out pupillary light reflex for the particular application of video summarization. The residual is converted into an exciting-ness score for each frame of a video. We show results on a variety of videos, and compare against ground truth as reported by three independent coders.

People often under or over-estimate their performance on interactive learning experiences with virtual agents, especially when it comes to interpersonal skills such as empathy. To generate more accurate self-assessment of performance, we propose the use of in-action reflective learning opportunities during interactive learning experiences with virtual agents. We conducted a user study in which third-year dental students (n=58) participated in an interactive learning experience that required them to demonstrate empathy towards a virtual agent playing the role of a patient. During the interaction, an in-action reflective learning intervention prompted the students to self-assess their performance with regards to empathy. Our results show that students' self-assessment correlates to an assessment performed by outside raters, and that students were significantly more empathetic on a second opportunity to demonstrate empathy to the virtual patient after having performed their reflection.

In a group setting, it is possible for attributes of one group member to indirectly affect how other group members are perceived. In this paper, we explore whether one group member's agency (e.g. if they are real or virtual) can indirectly affect behavior with other group members. We also consider whether variations in the agency of a group member directly affects behavior with that group member. To do so, we examined gaze behavior during a team training exercise, in which sixty-nine nurses worked with a surgeon and an anesthesiologist to prepare a simulated patient for surgery. The agency of the surgeon and the anesthesiologist were varied between conditions. Nurses' gaze behavior was coded using videos of their interactions. Agency was observed to directly affect behavior, such that participants spent more time gazing at virtual teammates than human teammates. However, participants continued to obey polite gaze norms with virtual teammates. In contrast, agency was not observed to indirectly affect gaze behavior. The presence of a second human did not affect participants' gaze behavior with virtual teammates.

This paper presents an application of the CASSM (Concept-based Analysis of Surface and Structural Misfits) framework to interactive machine learning for a bodily interaction domain. We developed software to enable end users to design full body interaction games involving interaction with a virtual character. The software used a machine learning algorithm to classify postures as based on examples provided by users. A longitudinal study showed that training the algorithm was straightforward, but that debugging errors was very challenging. A CASSM analysis showed that there were fundamental mismatches between the users concepts and the working of the learning system. This resulted in a new design in which aimed to better align both the learning algorithm and user interface with users' concepts. This work provides and example of how HCI methods can be applied to machine learning in order to improve its usability and provide new insights into its use.

A reflective practicum is a low-pressure, low-risk learning environment. In a reflective practicum a learner is educated in a professional practice and how to use reflection in the setting of that professional practice. An example of a low-pressure and low-risk learning environment is the use of embodied conversational agents (ECAs) in medicine to provide training for interviewing and diagnostic skills. However, such ECA experiences have not been used to teach how to use reflection in the setting of a professional practice. In this paper we present a framework that supports explicit reflective learning for ECA experiences. Using this framework, ECA experiences become a reflective practicum. This framework was applied to an ECA experience called the Neurological Examination Rehearsal Virtual Environment (NERVE), and created a sample experience called the NERVE Reflective Practicum (NERVE-RP). We conducted a user study in which second-year medical students (n = 76) used NERVE-RP and engaged in reflection based on the experience. The results of the user study show that students engage in valuable reflections during the experience including instances of critical reflection.

Thanks to the decreasing cost of whole-body sensing technology and its increasing reliability, there is an increasing interest in, and understanding of, the role played by body expressions as a powerful affective communication channel. The aim of this survey is to review the literature on affective body expression perception and recognition. One issue is whether there are universal aspects to affect expression perception and recognition models or if they are affected by human factors such as culture. Next, we discuss the difference between form and movement information as studies have shown that they are governed by separate pathways in the brain. We also review psychological studies that have investigated bodily configurations to evaluate if specific features can be identified that contribute to the recognition of specific affective states. The survey then turns to automatic affect recognition systems using body expressions as at least one input modality. The survey ends by raising open questions on data collecting, labeling, modeling, and setting benchmarks for comparing automatic recognition systems.

The conveyance and recognition of affect and emotion partially determine how people interact with others and how they carry out and perform in their day-to-day activities. Hence, it is becoming necessary to endow technology with the ability to recognize users' affective states to increase the technologies' effectiveness. This paper makes three contributions to this research area. First, we demonstrate recognition models that automatically recognize affective states and affective dimensions from non-acted body postures instead of acted postures. The scenario selected for the training and testing of the automatic recognition models is a body-movement-based video game. Second, when attributing affective labels and dimension levels to the postures represented as faceless avatars, the level of agreement for observers was above chance level. Finally, with the use of the labels and affective dimension levels assigned by the observers as ground truth and the observers' level of agreement as base rate, automatic recognition models grounded on low-level posture descriptions were built and tested for their ability to generalize to new observers and postures using random repeated subsampling validation. The automatic recognition models achieve recognition percentages comparable to the human base rates as hypothesized.

Conveyance and recognition of human emotion and affective expression is influenced by many factors, including culture. Within the user modeling field, it has become increasingly necessary to understand the role affect can play in personalizing interactive interfaces using embodied animated agents. However, little research within the computer science field aims at understanding cultural differences within this vein. Therefore, we conducted a study to evaluate if differences exist in the way various cultures perceive emotion from body posture. We used static posture images of affectively expressive avatars to conduct recognition experiments with subjects from three cultures. After analyzing the subjects' judgments using multivariate analysis, we grounded the identified differences into a set of low-level posture features. We then used Mixture Discriminant Analysis (MDA) and an unsupervised expectation maximization (EM) model to build separate cultural models for affective posture recognition. Our results could prove useful to aide designers in creating more effective affective avatars.

Connection Science, Vol. 15, No. 4, December 2003, 259–269 Page 259, line 6, reads: [email protected] It should read: [email protected] Taylor and Francis Ltd apologises for this error and for...

Studies on emotion are currently receiving a lot of attention. The importance of emotion in the development and support of intelligent and social behaviour has been highlighted by studies in psychology and neurology. Hence, the recognition of affective states has also become a critical feature in robot social development, with robots assumed to take on a role as social companion. In this paper, we address the issue of endowing robots with the ability to learn incrementally to recognize the affective state of their human partner by interpreting their gestural cues. We propose a model that can self-organize postural features into affective categories, and use contextual feedback from the partner to drive the learning process.