Intelligent Systems for Assisted Cognition AwardsThe U.S. Department of Health and Human Services and U.S. Department of Labor estimate the number of individuals in the United States who will require paid long-term care services (due to age and disabilities) will likely double from the 13 million in 2000, to 27 million people by the year 2050. Given that medical sensors and personal monitoring devices are getting smaller, less expensive and more readily available and deployable there is a logical reason to develop and study systems that can at least partially address this growing need through technological means. The ultimate goal of this research is to enable individuals to live longer, fuller, safer and more independent lives. Intelligent Systems for Assisted Cognition is an emerging field of research with the goal of improving the lives of the cognitively or sensory impaired. Although initially focused on assisting individuals with special needs, these applications, tools and devices have the potential to enable the development of technologies that may facilitate everyday life for everyone. This RFP awarded $300,000 (50,000 each, US$) total to the following research projects/groups: Intelligent Systems for Assisted Cognition RFP Award Recipients
Supporting Alzheimer’s Patients through Memory Augmentation
Anind Dey
Carnegie Mellon University, USA A common symptom of Alzheimer's disease is the loss of episodic memory, resulting in loss of sense of self of the sufferer, and increased burden on the caregiver. This has negative medical, financial and social impacts. We are developing and evaluating a memory prosthesis that uses contextual cues and automated techniques for determining what cues will help with memory recollection, and what order the cues should be presented to help with recollection. We will evaluate our system to show that it reduces caregiver burden and more appropriately helps a person with episodic memory impairment recall past experiences. 
Automatic Generation of Adaptive Directions for Wayfinding
Gaetano Borriello, Kurt Johnson, Dieter Fox
University of Washington, USA Our research group proposes to apply a combination of AI planning techniques and user studies to create a system that can generate tailored outdoor (door-to-door) walking directions for the cognitively impaired. Our approach to creating routes will involve choosing an optimal path that maximizes the expected likelihood that the user will successfully reach the destination. The path consists of connected nodes, with each having a set of possible directions that could be presented to the user. Using a Markov decision process (MDP) framework, we will research how to assign weights to path nodes. These weights will reflect probabilities that a user will succeed at making the correct wayfinding choice at that node given the best direction to give at that location, which could be a symbol, visible landmark, or some more sophisticated representation of where to go. In order to quantify the “quality” of a direction, we will conduct user studies to learn what features (e.g., landmark type or distance, etc.) are easiest for our intended users in different situations (e.g., degree of the node, location relative to roads or free space, etc.). Our system will be able to create routes that are adapted to each individual user, because weights can be adjusted based on individual preference and capability. 
An Online Community for Teachers to Support, Observe, Collect and Evaluate Assisted Communication with Children with Autistic Spectrum Disorder
Gondy Leroy, Claremont Graduate University, USA
Gianluca De Leo, Old Dominion University Teachers play an important role in the effort to increase the communication skills of children with autism and determine the required teaching style. However, they do not have any tools that allow them to share ideas, data, and information, and provide objective feedback on communicative behavior of children with autism. We are developing a communication system for severe autistic children using SmartPhones to enable on-the-go communication. The project for which we are seeking support will leverage our current research and develop and test the support environment that lets teachers, therapist and parents evaluate longitudinal communication data from the SmartPhones for individual children and for groups of children, discuss findings, and complement this data with other child-specific data using modern data analysis and data mining tools. If successful, this project will help teachers in their daily activities and that will also support and encourage integration of children with autism in our society. The project will make contributions in the field of special education and software development for users with special needs. 
Mixed-Initiative Visual Vocabulary Application for People with Aphasia
Perry Cook,
Princeton University, USA The concept of “average user” is even less applicable to a user population of individuals with aphasia, a language impairment, than to a non-disabled one. Current assistive communication technologies lack flexibility, opportunity for customization, and extensive vocabulary. As a step towards an effective communication aid and to address the lack of vocabulary depth and breadth, we propose a visual vocabulary application that is both adaptive and adaptable. The application uses as a framework a ubiquitous computerized lexicon that was created to mimic the systemic organization of language in a human mind, WordNet. It allows vocabulary personalization, intelligent vocabulary organization and management, and facilitates context-aware vocabulary search. The system will provide a list of words that frequently pair with a selected word based on the general word co-occurrence as well as the statistics data collected from the usage of the system; thus words relevant to the user’s interests or the context of the communication will surface faster. 
Behavioral Imaging Technologies to Support Early Detection of ASD
Gregory Abowd, James Rehg, Rosa Arriaga
Georgia Institute of Technology, USA We propose to develop new techniques for automated video analysis that can address these twin challenges of search and coding in the retrospective video studies for autism research, resulting in dramatic increases in speed and consistency. We expect our techniques to benefit both the quantitative analysis of diagnostic instruments obtained in a clinical setting (such as ADOS), as well as supporting more efficient and effective retrospective video studies. In addition, our work will popularize the analysis of infant social interactions within the activity recognition community, and encourage other computer science researchers to consider the challenges posed by ASD research. In the scope of this one-year grant, we will conduct a preliminary study that demonstrates the feasibility of automated visual analysis of infant social interactions (Peek-a-Boo in particular) and its use in both clinical diagnostic settings and retrospective video studies. 
A Software Framework for Domestic Cognitive Training Services
Amedeo Cesta
Institute for Cognitive Science and Technology; Italian National Research Council (ISTC-CNR), Italy Research in neuro-psychology and gerontology has demonstrated the beneficial effects of memory training on various forms of age-related cognitive impairment. It has also been shown that training is particularly effective if associated to ecological tasks, i.e., activities regarding people's every-day life and environments. We propose a novel use of technology aimed at incorporating cognitive training within the everyday life of the assisted user. Specifically, we put forth the idea of developing a “Cognitive Gym”, i.e., a software infrastructure for integrating current state of the art AAL, AI and user-interface technology to provide cognitive training in the context of domestic activities. The idea is grounded on two key observations: (a) cognitive training programs are maximally effective if they occur in an ecological context; (b) a major obstacle to overcome in order to develop adaptive and highly-contextualized intelligent assistants for cognitive training is the issue of technology integration. Intelligent Systems for Assisted Cognition RFP
|
