Opportunity Description:

Purdue has launched an ambitious Integrative Data Science Initiative focused on advancing the frontiers of research and the application of data science to pressing, socially relevant issues as well as a new campus-wide, transformational data science education initiatives. The initiative is designed to build on and advance Purdue’s existing strengths to position the University as a leader at the forefront of advancing data science-enabled research and education. 

This RFP represents the first investment towards achieving the goals of the Integrative Data Science Initiative. For this first round, the areas of focus will be health care; defense; ethics, society and policy; fundamentals, methods and algorithms; and cross-cutting data science-enabled research. Discovery Park intends to fund several projects at the level of $100K to $150K per year for two years. We encourage high-risk, high-payoff proposals that describe "game changing, transformative ideas" (not extensions of existing research) targeted to one of the four initiative focus areas.

• Data Science for Health Care: This focus area is for proposals that have the potential to significantly transform health and health care through innovative data science-enabled research.  Examples include, but are certainly not limited to, novel approaches to generate evidence from a variety of data sources for personalized care; data integration with privacy preserving analytics, characterizing social and behavioral circumstances that effect health outcomes, data science enabled care transitions, and analytics to support integrated care and population health.  Proposals should bring together researchers across multiple disciplines, be truly transformative in scope, and have the opportunity to be sustainable after the 2-year funding period ends. 

• Data Science for Defense: Concepts are sought that could dramatically improve the efficiency of data science in accomplishing defense and security missions. One example category is "Designing the Data"- how to optimize data gathering and structuring. In a sparse data setting, what is the most valuable data to collect and/or combine with synthetic data? In a big data setting, what should be ignored and what should be fused? In any setting, how to form models of trust in the data used for critical decisions and how can (readily available) non-defense/intel/security data sets serve as useful surrogates to test & validate algorithms for defense applications. Another example category is "Optimal Sensemaking"- the maximization of human-machine symbiosis via combinations of novel data representations, algorithms, visualizations, and computing architecture that, for example, enable never-before see functionality on small, edge devices and platforms (e.g., small unmanned aircraft, unintended ground sensor network node, satellite, etc.). 

• Data Science for Ethics, Society and Policy: Proposals in this focus area bring social scientific and humanistic inquiry together with data science. This can include projects that are fundamentally translational or methodological. For example, a project might “scale up” a traditional social science or humanistic approach for application to big data. Projects could bring data scientists into interdisciplinary conversations about eliminating bias, improving transparency or balancing incommensurables in designing algorithms. Social science and humanistic applications of data science may also foreground methodological questions, such as how to diagnose and avoid analytic errors. The call is open as to the substantive issue to which data science might apply.  

• Data Science for Fundamentals, Methods and Algorithms: This focus area is for proposals that elicit a better understanding of core concepts in data science including: information, knowledge, fairness, trust, risk, collusion, privacy, and information-efficient computation.  Research topics might include characterization of analytical bounds (limits) on important parameters such as statistical significance, limits of learning, transferability and generalizability, control, and computation complexity; new methodologies for validations and verification on real data taking into account tradeoffs between computation and accuracy; and novel applications that rely on adversarial learning. Applicants are encouraged to incorporate other focus areas identified in this RFP in a multidisciplinary approach. 

• In addition, proposals are also requested that exploit synergies amongst the four areas above and where researchers from across disciplines will work together across these boundaries. Examples might include proposals exploring a problem at the nexus of policy and defense-related data science, or the application of data science fundamentals/theory to a health-related grand challenge, etc. 

The Proposal Process:

Teams must be composed only of Purdue faculty/staff and students, and must be led by a faculty member. We encourage early-career faculty to participate in this process both as team leaders and team members. 

A proposal narrative, not to exceed two pages (single spaced, 12-point font, Times Roman), should be submitted first. This proposal narrative should include:

• A brief executive summary that states the data science focus area and articulates the “game changing, transformative idea”, and the intellectual value or contribution represented by the idea.
• A brief description of the need that the idea tackles.
  • What is the un-met need that the idea addresses?
  • What are the goals, expected outcomes and impact of the project?
  • Why now?
  • A brief description of the team’s approach to meet the goals, including an articulation of the path to sustained external funding and how the approach leverages Purdue resources.
  • A brief description of the benefits of the proposed solution to Purdue.
  • Milestones that will be met over the period of performance (not to exceed 24 months).

Other considerations include the following:

• The proposal must clearly identify a team leader and project champion who must be a faculty member.
• The milestones should set targets for tangible progress to be achieved and will be reviewed every six months for continued funding. 

Budget

The program will fund several teams for up to two years (depending on scope of the proposal) with a pool of strategic investment resources. A detailed budget request aligned with the milestones should be included in the proposal but will not count toward the two-page limit. It is expected that a typical project will have a budget of up to $150K per year. However, final budget determinations will be made after selection based on final scope and schedule. 

Budgets must be approved by Sponsored Programs Services – Pre-award prior to proposal submission. There is no need to show cost-share in the proposals. Selected projects must not only have a rational plan to spend the resources, but actually do so; if money is not spent well in year 1, year 2 is not guaranteed. Full budget allocation to the project will be contingent on satisfactory progress in meeting the bi-annual milestones.

Proposal Evaluation

All proposals will be evaluated by a committee made up of the Directors from the Regenstrief Center for Healthcare Engineering (RCHE), Institute for Global Security and Defense Innovation (i-GSDI), Purdue Policy Research Institute (PPRI), and Center for the Science of Information (CSoI). This committee will select proposals for advancement to an oral presentation phase. 

Oral presentations will be evaluated by the same committee members and additionally include the Executive Director and Chief Scientist of Discovery Park, and two to three senior faculty members. Presentations will be strictly limited to 8 minutes with 7 minutes for questions. These presentations are expected to succinctly but clearly address all elements of the project, as well as describe the project team leadership qualifications, milestones and the budget. Details on presentation framework will be shared with the selected teams.

The proposals will be scored based on the following weighed criteria:

• 60% - Is this a new idea, or an extension of ongoing work? How transformative is the idea? Is the approach high-risk, high-payoff? How strong is the scholarship/research base for the proposal?
• 20% - How likely is it that the project funding will result in new federal, corporate, donor and/or foundation funding at significant scale? Identify likely funding sources as specific as possible.
• 20% - How strong is the track record of the team in successfully tackling complex problems and working collaboratively over the long term? Is the team interdisciplinary? 

Note: Care will be exercised to ensure that there is no conflict of interest in the proposal evaluation process.

Schedule

• Submission Deadline: 5:00 pm EST, April 19, 2018
• Notification for selection to present oral presentation: April 26, 2018
• Oral presentations: May 3, 2018
• Notification of awards: May 14, 2018
• Projects kickoff: May 29, 2018
• Project Period: May 29, 2018 – May 28, 2020 

Proposals will be due to the Discovery Park office by April 19, 2018. Submit all materials to IDSIatDP@purdue.edu.  

Additional Questions

Proposal procedures and general questions can be addressed to:

• Cliff Wojtalewicz, Assistant Director of Discovery Park: cliffw@purdue.edu. 

Budget-related questions should be directed to:

• Donna Brown, Assistant Director of Financial Affairs: brown746@purdue.edu 

Specific research-related questions can be addressed to:

• Data Science for Ethics, Society and Policy: Laurel Weldon, Director, Purdue Policy Research Institute, Distinguished Professor of Political Science at weldons@purdue.edu
• Data Science for Health Care: Paul Griffin, Director, Regenstrief Center for Healthcare Engineering, the St. Vincent Health Chair of Healthcare Engineering and Professor of Industrial Engineering at paulgriffin@purdue.edu
• Data Science Fundamentals, Methods and Algorithms: Wojciech Szpankowski, Director, Center for the Science of Information and the Saul Rosen Distinguished Professor of Computer Science at szpan@purdue.edu