1. The Mycroft Cognitive Assistant: Enhancing and supporting T-shaped skills and research collaborations
Kara L. Hall, PhD
Director, Science of Team Science (SciTS), Division of Cancer Control and Population Sciences, National Cancer Institute
Paul K. Courtney
Chief Technology Officer, Open Health Systems Laboratory
Anil Srivastava,
President, Open Health Systems Laboratory

2. The Vision…
A scientific enterprise where cognitive mediators regularly serve as an integral part of a research team throughout the entire research process. 

3. Overview Define team science and the science of team science
Describe a four-phase model of research, including key team processes.
Highlight a new initiative aimed at producing a suite of tools designed to enhance the effectiveness and efficiency of research collaboration.  
Outline potential functionality to support each research phase using IBM Watson platform for building the Mycroft system

4. Team Science and The Four-Phase Model of Transdisciplinary Research

5. What are we trying to do? NIH Mission
….enhance health, lengthen life, and reduce illness and disability
Social work - improve the quality of life and enhance wellbeing of the nation;
Social work - helping profession; the main goal of social work is to improve a society’s overall well-being, especially for the most vulnerable populations.

6. Why team science? The societal and scientific problems are complex –
Multi-level, multi-factorial, interacting influences.

7. Variations in Team Science

8. Dimensions of Team Science that create unique profiles and challenges
Range
Diversity
Homogeneous
Heterogeneous
Integration
Unidisciplinary
Transdisciplinary
Size
Small (2)
Mega (1000s)
Proximity
Co-located
Globally distributed
Goal alignment
Aligned
Divergent or Misaligned
Boundaries
Stable
Fluid
Task interdependence
Low
High

9. Adapted from: Rosenfield, 1992; Stokols, Hall, et al
Adapted from: Rosenfield, 1992; Stokols, Hall, et al., 2008 Falk-Krzesinski, H.J., 2012; Austin et al., 2008; Nissani, 1995

10. Collaboration is complex Multi-level contextual factors
Intrapersonal
Members' attitudes toward collaboration and their willingness to devote substantial time and effort to TD activities
Members' preparation for the complexities and tensions inherent in TD collaboration
Participatory, inclusive, and empowering leadership styles
Physical Environmental
Spatial proximity of team members' workspaces to encourage frequent contact and informal communication
Access to comfortable meeting areas for group discussion and brainstorming
Availability of distraction-free work spaces for individualized tasks requiring concentration or confidentiality
Environmental resources to facilitate members' regulation of visual and auditory privacy
Societal/Political
Cooperative international policies that facilitate exchanges of scientific information and TD collaboration
Environmental and public health crises that prompt inter-sectoral and international TD collaboration in scientific research and training
Enactment of policies and protocols to support successful TD collaborations (e.g., those ensuring ethical scientific conduct, management of intellectual property ownership and licensing)
Organizational
Presence of strong organizational incentives to support collaborative teamwork
Non-hierarchical organizational structures to facilitate team autonomy and participatory goal setting
Breadth of disciplinary perspectives represented within the collaborative team or organization
Organizational climate of sharing
Frequent opportunities for face-to-face communication and informal information exchange
Technological
Technological infrastructure readiness
Members' technological readiness
Provisions for high level data security, privacy, rapid access and retrieval
Interpersonal
Members' familiarity, informality, and social cohesiveness
Diversity of members' perspectives and abilities
Ability of members to adapt flexibly to changing task requirements and environmental demands
Regular and effective communication among members to develop common ground and consensus about shared goals
Establishment of an hospitable conversational space through mutual respect among team members
Collaborative Effectiveness of Transdisciplinary Science Initiatives
Stokols, Misra, Hall, Taylor, & Moser, 2008

11. The Science of Team Science (SciTS) is a cross-disciplinary field of study that aims to: (1) build an evidence-base and (2) develop translational applications to help maximize the efficiency and effectiveness of team-based research.

12. SciTSJournal Supplement
Building the SciTS Evidence-base
NCI Conference: The Science of Team Science: Assessing the Value of Transdisciplinary Research
Applying the Science of Teams to inform Policy and Research on Team Science
Mapping a Research Agenda for SciTS
Annual International
SciTS Conference
2006
2008
2010
2011
2012
2013
2014
2015
Team Approaches to Science, Practice, and Policy in Health
SciTSJournal Supplement
Collaboration Science and Translational Medicine
National Academies
Consensus Study

13. Developing Translational Applications

14. Four-Phase Model

15. Note: Gray lines represent potential alternative, iterative, or recursive pathways
Hall, KL, Vogel, AL, Stipelman, B, Stokols, D, Morgan, G, & Gehlert, S. (2012). A four-phase model of transdisciplinary research: goals, processes and strategies. Translational Behavioral Medicine, 2, 4, 15

16. Development Phase: Goals and Key Processes
Goal: Define the scientific or societal problem space of interest, including identifying the intricacies and interconnections of concepts that fall within the problem space and establishing the boundaries of the problem space to be addressed.
Key Processes: Encourage information sharing and integrative knowledge creation among diverse participants
Generate shared mission and goals
Develop critical awareness
Externalize group cognition
Developing group environment of psychological safety
Team Type:
Network, working group, advisory group, emerging team
Engage in a group process to define a TD problem space by collaboratively generating a cognitive artifact that helps to articulate the complexities of the problem space and the wide variety of relevant disciplines and fields.
Hall, KL, Vogel, AL, Stipelman, B, Stokols, D, Morgan, G, & Gehlert, S. (2012). A Four-Phase Model of Transdisciplinary Research : Goals, Processes and Strategies. Translational Behavioral Medicine, 2 (4). 16

17. Conceptualization Phase: Goals and Key Processes
Goal: Develop novel research questions, hypotheses, a conceptual framework, and a research design that integrate collaborators’ disciplinary perspectives and knowledge domains to address the target problem in innovative ways.
Key Processes: Facilitate integrative knowledge creation among team members and the development of a research plan
Create shared mental models
Generate shared language
Develop compilational transactive memory
Develop team TD ethic
Team Type:
Emerging team, evolving team
Approaches and Activities:
The group hosted monthly seminars for core group members and potential future collaborators.
Each seminar began with a section introducing basic disciplinary concepts to foster understanding of the expertise of each presenter (compilational transactive memory), become familiar with the terminology and concepts in each other’s disciplines (shared language) and develop an appreciation for the contributions of each discipline to the shared goals (team TD ethic).
In addition, glossaries and reading lists were developed to generate shared language.
The group developed a shared mental model which was reflected in the development of a conceptual model of specific research goals
Developed a conceptual model that identified the potential contributions of disparate levels of science. A subset of investigators moved forward to develop TD hypotheses regarding the neural mechanisms that underlie medication effects on early nicotine abstinence symptoms.
Lerman, 2012
Use public seminars among collaborators to help develop compilational transactive memory, shared language for a TD research collaboration, team TD ethic, and shared mental model of research collaboration.
Hall, KL, Vogel, AL, Stipelman, B, Stokols, D, Morgan, G, & Gehlert, S. (2012). A Four-Phase Model of Transdisciplinary Research : Goals, Processes and Strategies. Translational Behavioral Medicine, 2 (4). 17

18. Implementation Phase: Goal and Key Processes
Goal: Launch, conduct, and refine the planned TD research
Key Processes:
Developing a shared understanding of… who knows what (compilational), who does what (compositional), how things get done (taskwork), and how interactions occur among the research team (teamwork)
Compositional, Taskwork, and Teamwork Transactive Memory
Conflict Management
Team Learning (e.g., reflection, action, feedback, discussion)
Team Type:
Real team
“Real” vs “Psuedo” team
Characteristics that lead to increased performance and innovation
Interdependent
Iterative reflection (systematic consideration of team performance and participation in related adaptation to team goals and processes)
Demonstrate clear understanding of team membership
West et al, 2011; West & Lyubovikova, 2012
Hall, KL, Vogel, AL, Stipelman, B, Stokols, D, Morgan, G, & Gehlert, S. (2012). A Four-Phase Model of Transdisciplinary Research : Goals, Processes and Strategies. Translational Behavioral Medicine, 2 (4). 18

19. Translation Phase: Goals and Key Processes
Goal: Apply research findings to advance progress along the discovery–development–delivery pathway to ultimately provide innovative solutions to real-world problems.
Key Processes:
The evolution of the team, as needed, to identify and pursue translational goals
Develop shared goals for the translational endeavor
Develop shared understandings of how these goals will be pursued
Team Type:
Adapted team, new team
Initiate community outreach activities to identify translational partners to evolve the TD team. Work together to identify and implement translational goals in ways that draw upon the expertise of both investigators and translational partners.
Hall, KL, Vogel, AL, Stipelman, B, Stokols, D, Morgan, G, & Gehlert, S. (2012). A Four-Phase Model of Transdisciplinary Research : Goals, Processes and Strategies. Translational Behavioral Medicine, 2 (4). 19

20. What is Cognitive Computing?
Cognitive computing is a relatively new kind of computing. This technology can be used to understand and find patterns in human-friendly unstructured data such as research papers, articles and even novels. Unstructured data contains grammar, ambiguity and often requires a context to understand compared to structured data that is based on a pre-defined data model. 1 Cognitive computing is a new type of software that combines machine learning and natural language processing techniques in order to solve complex problems. IBM Watson is a revolutionary system implementing the cognitive computing paradigm.2
1 A Cognitive Assistant Powered by IBM Watson: A Tool for Finding Researchers Within the Medical Field. Pontus Thome, Erik Sievers, Hannes Haggander, Rasmus Letterkrantz, Fredrik Rahn. Chalmers. Gothenberg, Sweden. January 2016.
2http://technobium.com/getting-started-with-ibm-watson/ Accessed 3/11/2016

21. What is a Cognitive Assistant?
A software agent (cog) that
“augments human intelligence” (Engelbart’s definition1 in 1962)
Performs tasks and offer services (assists human in decision making and taking actions)
Complements human by offering capabilities that is beyond the ordinary power and reach of human (intelligence amplification)
1Augmenting Human Intellect: A Conceptual Framework, by Douglas C. Engelbart, October 1962
From Cognitive Assistance at Work: Cognitive Assistant for Employees and Citizens, by Hamid R. Motahari-Nezhad, AAAI 2015 Fall Symposium.

22. Why do we need Cognitive Assistants?
“The reason I was interested in interactive computing, even before we knew what that might mean, arose from this conviction that we would be able to solve really difficult problems only through using computers to extend the capability of people to collect information, create knowledge, manipulate and share it, and then to put that knowledge to work…Computers most radically and usefully extend our capabilities when they extend our ability to collaborate to solve problems beyond the compass of any single human mind.1”
1 Improving Our Ability to Improve: A Call for Investment in a New Future. Douglas C. Engelbart, September 2003.

23. The Mycroft Cognitive Assistant®
Mission: Provide a suite of tools designed to enhance the effectiveness and efficiency of biomedical research collaboration. Deliver solutions that support and facilitate collaboration across entire the research process, including exploring a research problems space, assembling research teams, to obtaining research funding, launching a research project, interacting with team members, enhancing team functioning, as well as developing and disseminating research products. IBM Watson cognitive computing technology drives the intelligence of the tools to provide a dynamic set of cognitive assistants to optimize research productivity and increase innovative capacity. We have named our development cog “Sherlock”

24. Sherlock High-Level Project Goals
Assist in the assembly of research teams that includes the right expertise regardless of the type of knowledge or location of collaborators.
Match researchers to appropriate funding opportunities based on research interests, expertise, skills, type of research proposal, and the grants’ research objectives.

25. Unstructured data sources
Social Media
CVs and Biosketchs
Funding Announcements
Publications

26. Preliminary Taxonomy Map
50 coded empirical studies on research teams
300+ constructs and 500+ relationships among them
4 content domains: individual, team, institutional/organizational, policy/societal
5 levels of granularity (4 displayed):
Huang, G., Dathe, M., Stipelman, B. A., Gibbs, K. D., Serrano, K. J., Vogel, A. L., Larsen, N., Williams, C., Tsakralides, S. P., Hammond, R., &. Hall, K. L. (2015, June). Developing a systems map of team science: A spotlight on methods and preliminary results. Poster presented at Sixth Annual International Science of Team Science Conference, Bethesda, MD. (the poster is accessible here: )

27. Matching Four Phase Model Objectives with Watson Capabilities
Watson cognitive assistance
Development
Encourage information sharing and integrative knowledge creation among diverse participants.
Develop group environment of psychological safety through the use of Sentiment Analysis & Personality Insights of communication in immersive environments.
Conceptualization
Facilitate integrative knowledge creation among team members and the development of a research plan.
Develop compilational transactive memory – or a knowledgebase of “who knows what” using Natural Language Classification of individual’s publications, blogs, social media & professional engagement.
Implementation
Developing a shared understanding of who knows what (compilational), who does what (compositional), how things get done (taskwork), and how interactions occur among the research team (teamwork).
Iterative reflection (systematic consideration of team performance and participation in related adaptation to team goals and processes) through Watson-augmented metacognition by team members on team performance
Translation
Support the evolution of the team, as needed, to identify and pursue translational goals.
Identify individuals who can help with expertise in the interventions or in the community through the use of Watson-enabled research networking.
Adapted from: Hall, K. L., Vogel, A. L., Stipelman, B., Stokols D., Morgan, G., & Gelhert, S. (2012). A four-phase model of transdisciplinary research: goals, team processes, and strategies. Translational Behavioral Medicine, 2 (4),

28. Watson Augments T-Shaped Skills
Accessed 3/12/2016

29. Our Cognitive Computing Partners

30. Thank you