Abstract: Research and practice at the intersection of artificial intelligence (AI) and global health has grown rapidly in the last few years, yet most of these efforts are fragmented and disconnected. There has been a pressing need to facilitate knowledge-sharing and resource coordination in this space across research and practice and multiple disciplines. With support from the IPaT engagement grant, we launched a global, interdisciplinary Research to Impact Collaborative (RIC) on AI and global health that: 1) supports knowledge-sharing across research and practice, 2) facilitates student learning, and 3) accelerates cross-sector collaborations. To catalyze the RIC, we have been conducting a year-long virtual seminar series and multiple in-person workshops that bring together researchers, practitioners, and students. This talk will reflect on the process of building a lasting collaborative, and laying the foundation for interdisciplinary collaborations and future funding.

Bio: Naveena Karusala is an Assistant Professor in Interactive Computing at Georgia Tech. Her research is on human-centered AI and health/wellbeing in underserved communities. She examines how we can design for more just care infrastructures that better recognize the labor, autonomy, and rights of care workers and care recipients. Her work has received Best Paper at ACM CHI and Diversity and Inclusion Recognition at ACM CSCW. She was previously a postdoctoral fellow at the Harvard Center for Research on Computation and Society, and holds a PhD in Computer Science from the University of Washington. Naveena also serves as Vice-President for Communications on the ACM SIGCHI Executive Committee.

- - - - - - - - - -

Abstract: Data centers are growing rapidly in the U.S., with nowhere more notable than in Georgia, particularly in the Atlanta metropolitan region. Moreover, data centers are often sited in ways that impose local external costs, impacting important aspects of everyday life, such as water security, energy prices, taxes, jobs, housing, and air quality. This talk shows how design-based workshops can be tools to investigate these trade-offs in constructing data centers, weighing the economic benefits against their external impacts on local Atlanta communities. Our talk in particular will discuss the process of designing these workshops, and the proposed outputs.

Bio: Cindy Lin is the Stephen Fleming Early Career Assistant Professor at the School of Interactive Computing at Georgia Tech. She is the director of Critical Technocultures Lab. Her research interests include data politics in environmental governance in Southeast Asia and the United States, and, more recently, rentier relations in global data center infrastructures. 

Allen Hyde is an Associate Professor and Director of Undergraduate Studies in the School of History and Sociology at the Georgia Institute of Technology. He is a quantitative scholar whose main research areas are stratification and inequality, urban sociology, work and occupations, climate and disaster resilience, and immigration. He is currently researching the effects of race/ethnicity and immigration status on homeownership, social and demographic change in Clarkston, GA (known as the most diverse square mile in America), and was recently Principal Investigator for the Youth Advocacy for Resilience to Disasters Program research project funded by the National Science Foundation's Civic Innovation Challenge.

- - - - - - - - - -

Abstract:  We developed multimodal systems that transform the movement of the hand, arm, trunk, and head into the sounds of musical instruments (i.e., sonification). In one system, a depth camera captures human motion and creates associated musical sounds depending on the movement characteristics of the hand, arm, trunk, and head. The potential use cases include the rehabilitation training of reaching and grasping in stroke survivors and maintaining the head position upright in cerebral palsy. In another system, a virtual reality (VR) headset captures hand trajectory of the user in three dimensions in the VR environment. When the hand traces the visually presented target path appropriately, musical melody is produced. This modality could be used for training the arm and shoulder movements in people with limited range of motion, for example. This type of transformation of movement trajectories into musical sounds has the potential to empower people with motor disabilities to express their artistry with their less-impaired body parts as well. It could be personalized and is expected to benefit older adults and individuals with motor impairments by enhancing their well-being by introducing new, enjoyable, engaging, and rewarding artistic expressions or exercises.

Bio: Dr. Minoru “Shino” Shinohara is the director of the Human Neuromuscular Physiology Lab and an associate professor at the School of Biological Sciences at Georgia Tech. He received his PhD in Multidisciplinary Sciences (Life Sciences) from the University of Tokyo, Japan. His research focuses on neurophysiological and biomechanical mechanisms that underpin motor skills, with a particular emphasis on their adaptations to altered neural input, aging-related changes, practice, and rehabilitation in humans. He employs state-of-the-art techniques in neuroscience, neuroengineering, physiology, and biomechanics, including transcranial magnetic stimulation and transcutaneous vagus nerve stimulation. In addition, he collaborates with engineers in the fields of human-robot interaction, human augmentation, machine learning, and their application in the enhancement of physical activity, sports, health, and rehabilitation.

- - - - - - - - - -

Abstract:  As robots increasingly operate within public spaces and urban environments, they are frequently confronted with morally ambiguous situations that are dynamic, complex, and unpredictable. Such contexts introduce novel factors and forms of agency that may quickly exceed the scope of anticipated—or pre-programmed—human–robot interactions (HRI). This raises a critical question: can morally appropriate robot behaviors be generated for unforeseen HRI scenarios using large language models (LLMs), and if so, how? In this talk, I present our proposed framework for generating morally appropriate robot behaviors, illustrated through the exemplar of a mobile, space-making robotic reading companion. I also briefly report on our current progress in evaluating this framework. Broadly, the framework comprises three key steps: (1) defining a library of robot behaviors; (2) training LLMs on general moral principles alongside examples of morally appropriate behaviors in predefined scenarios; and (3) leveraging the LLM to generate candidate behaviors for novel situations. The generated behaviors are subsequently evaluated by users and relevant stakeholders with respect to their perceived moral appropriateness. Through this work, we aim to offer a practical approach for designers and roboticists to develop novel robotic systems capable of behaving ethically in everyday, unforeseen contexts.

Bio: Yixiao Wang is an Assistant Professor at the School of Industrial Design (SID), Georgia Institute of Technology (Georgia Tech). His background is in Architecture (B. Arch from Hunan University and M. Arch from UC Berkeley) and Human-Centered Design (Ph.D. from Cornell University). As an interaction designer and researcher, his work focuses on the human-centered design, engineering, and evaluation of “Space Agents” – physical and virtual spaces that are socially intelligent and interactive, as if they were our friends, partners, or companions. He is particularly interested in the various applications “Space Agents” could have in our everyday life, together with their lasting societal, cultural, and ethical values in education, healthcare, and work life. His major research fields are Architectural Robotics, Human-Robot Interaction (HRI), and Human-Agent Interaction (HAI), broadly situated within Human-Computer Interaction (HCI). He is the director of the “Robotic Environment Lab” at Georgia Tech, SID.