Virtual Patients Group - Improving
communication skills education using virtual patients.
|
Technical Innovations |
|||
|
|
|
Interpersonal
Simulator (IPS) (Developers: Johnsen, Raij,
Dickerson, Kotranza, Rossen, Chuah) The interpersonal simulator system
enables students to practice communication skills with virtual humans. The system supports many interface
modalities. IPS
is a software system that enables users to practice an interpersonal scenario
(typically a conversation) with a virtual human. Input: the user interacts with virtual
humans using a combination of: typing natural language, natural language
speech, multiple-choice, gestures, touch, and devices such as a Nintendo
Wiimote. The user is tracked using cameras and retro-reflective markers. Conversational Modeling: Crowdsourcing approach to virtual
human creation enables scenarios to be modeled in weeks. Cognitive: Currently we employ a keyword
matching approach with a statistical model.
More complex models can be employed. Output: The virtual human can be presented
as any of the following: a life-sized 3D character using a HMDs, projectors,
or large-screen TVs, as an avatar in SecondLife or a web-browser video
chat-bot, or as an online text chat-pot. Demos:
Interview
a pharmacy patient online, Video
chat with a pharmacy online Videos:
[Overview]
Publications: [Validity][IPS] Funded
by the National Science Foundation. |
|
|
|
|
Mixed Reality
Humans (MRHs)* (Developer: Kotranza) Integrated
virtual humans (to practice communication skills) with active and passive
haptics (to practice procedure). The
MRH platform combines IPS with both passive and active haptic systems. The passive haptics are used to enhance the
realism and guide the user’s touch.
The active haptics are integrated with the virtual human such that
they can respond to the user’s touch.
For example, if the user presses too hard, the sensors in the physical
simulator measurers the pressure and passes that to IPS. Then, IPS has the virtual human say that
they are in pain and present the appropriate facial animation. MRHs
have been shown to enhance user’s empathy to the VP, and it has been tested
with hundreds of health profession students. Videos:
[Clinical
Breast Exam] [Bi-directional
Touch] Publications: [Acceptability and
Usability of MRHs] [Bi-directional
Touch] IEEE Virtual Reality 2008
Best Paper Funded
by the National Science Foundation. |
|
|
|
|
Virtual Multi-tools
(NERVE) (Developer: Kotranza, Horton, Hwang)
Integrated commodity interfaces to provide robust and interactive experiences
with virtual humans with rarely-seen conditions. The
IPS system has been integrated with commodity interfaces, specifically the
Nintendo Wiimote. The Wiimote enables
the user to have a robust mechanism to gesture and interact with the user. For example in the eye-exam scenario, the
Wiimote represents the user’s hand, enabling them to robustly raise and lower
fingers, move their hands about, use virtual tools such as an ophthalmoscope,
and administer neurological tests. Studies
showed that providing haptic interfaces enabled the system to mask errors
with interfaces, such as speech recognition and understanding. Videos:
[Cranial
Nerve 3 Exam] Publications: [Virtual
Multi-tools] Funded
by the National Institutes of Health. |
|
|
|
|
Interpersonal
Visualization (IPSViz)* (Developer: Raij) Developed
technologies that enable students and educators to reflect and analyze their
interactions with virtual humans. IPSViz
– Interpersonal Visualizer is a software system that takes IPS interactions
and analyzes and visualizes the log files.
Two separate systems have been created: IPSViz^1 – This software takes an individual
IPS log file and enables reviewing the interaction from different
perspective, highlights conversational topics visited and avoided, and
presents the interactions of experts as a comparison for self-reflection. IPSViz^n – This software enables exploration
of multiple IPS log files, for example, educators identifying trends and
outliers in a class wide use of IPS. Online
demos: IPViz^1 (id/pass:
10157/00000) and IPSViz^n.
Video: [IPSViz]
Publications: [IPSViz] Funded
by the National Science Foundation. |
|
|
|
|
Virtual Social
Perspective-Taking (VSP) (Developers: Raij, Kotranza)
Developed interactions where the user re-experiences an interaction with a
virtual human from the perspective of the virtual human – in essence,
interacting with oneself. The
VSP system enables users to immediately revisit their IPS interaction with a
VH, but from the perspective of the VH.
Recorded video, head tracking data, and speech data are used to have
the user play the role of the VH talking with a recorded version of
themselves. This system enables users
to interactively evaluate their affect, observe non-verbal behavior, and
identify conversational opportunities for empathy. |
|
|
|
|
Distributed
Conversational Modeling (Virtual People Factory)* (Developers: Rossen) VPF is software
that enables content-area experts (e.g. educators and end-users) to rapidly
create virtual humans. Leveraging
crowdsourcing, VPF has end-users and educators model conversations between
users and VHs. Using a web-browser
interface, scripts can be created by leveraging a few educators and many
students to model the conversation scenario.
As
opposed to the common model of having experts and VH developers model scripts
used for training, the fundamental realization is that experts can not
anticipate the varied approaches that novices will try to address a topic. Thus it is critical to involve both
educators and students in the conversational modeling process. This
results in new virtual human conversation scenarios being conversationally
modeled in approximately two weeks as opposed to 6 months+ in pervious
methods. Further the VPF created
conversation models are more significantly accurate and contain more detailed
responses by the VH. Papers:
[XXX], Videos [XXX], Demos: Interview
a pharmacy patient online |
|
|
|
|
Virtual Patients +
Human Patient Simulators* (Developers: Hwang) Integrated
METI’s Human Patient Simulator (HPS) with IPS to enable hands-on procedure,
interaction, and communication skills training. IPS
has been integrated with the Human Patient Simulator (HPS) via its HIDEP
protocol. The VP is aware of the HP’s
vital signs, such as blood pressure, heart rate, and respiration. This integration leverages the VP’s ability
to 1) handle interpersonal communication, 2) present a variety of virtual
human models (weight, ethnicity, and gender), and 3) present dynamically visuals
difficult to show with physical devices (e.g. lips turning blue, vomit,
burns, the state of internal components). This
integration of IPS and HPS creates experiences that completely simulate a
variety of complex verbal+haptic scenarios that focus on procedure and
communicate affect. Videos:
[VP
communicating with HPS] Publications: [in submission] |
|
|
|
|
Bias with Virtual
Humans (Developers: Rossen) Study to the degree that virtual
humans elicit bias and can be used to educate users on cultural competency
skills. We
are exploring if real-world biases transfer to the virtual world. That is, do virtual humans that present
different characteristics, such as ethnicities, gender, age, and weight,
cause users to treat them similarly to the way they would treat real people with
similar backgrounds? To study this, we
have worked to correlate the Implicit Association Test (IAT) with video coder
reviews of medical students interviewing and empathizing with virtual human
patients. Our work has highlight the
correlation of IAT skin-tone bias to empathy exhibited to VHs of varying
skin-tones. Videos:
[XXX] Publications: [Skin-tone bias with
VHs] Funded
by the National Science Foundation. |
|