The Michigan Institute for Computational Discovery ... · Graduate Certificate in CDE Graduate Certificate in Computational Neuroscience Ph.D. in Scientific Computing. 4 What is MICDE?

The Michigan Institute for Computational Discovery & Engineering

2019 Catalyst GrantsInformational Session

February 7, 2019

2

What is MICDE?

The Michigan Institute for Computational Discovery and Engineering (MICDE) focuses on the development and innovative use of mathematical algorithms and

models on high performance computers (HPC) to support basic and applied research and development across a wide spectrum of disciplines in science and

engineering.

Over 140 affiliated faculty from 31 different departments

3

What is MICDE?



engineering.

Over 140 affiliated faculty from 31 different departments

Graduate Certificate in CDEGraduate Certificate in Computational NeurosciencePh.D. in Scientific Computing

4

What is MICDE?



engineering.

Seminar series, Symposia, Faculty Workshops


5

What is MICDE?



engineering.



Outreach and Industrial Engagement

6

What is MICDE?



engineering.


Three Research Centers

Graduate Certificate in CDEGraduate Certificate in Computational NeurosciencePh.D. in Scientific ComputingOutreach and Industrial

Engagement

MICDE is a unit of Advance Research Computing

8

Successful Case: NSF MRISeptember 2015

Center for Data-Driven Computational Physics

(A result from 2 faculty workshops in summer 2014…)• PI – Karthik Duraisamy (Aero) • $3.5M NSF Major Research Instrumentation (MRI): ConFlux

• Climate systems interactions• Cosmology• Computational materials physics• Turbulence in fluid flow• Subject-specific blood flow modeling

9

Successful Case: NSF DIBBsSeptember 2015

• Center for Network and Storage-Enabled Collaborative Computational Sciences• PI – Shawn McKee (Physics) • $5M NSF Data Infrastructure Building Blocks (DIBBs): MI-OSiRIS

• Data-driven analysis of genetic and molecular disease mechanisms• Simulated ocean floor• Biomedical modeling• Cosmology

10

Successful Case: NSF CRISPSeptember 2016

Interdependencies in Community Resilience: A Simulation Framework

• PIs – Sherif El Tawil (CEE)• $2.2M NSF Critical Resilient Interdependent Infrastructure and

Processes (CRISP)• MICDE provided a letter detailing institutional support: educational programs, research, ARC• CSCAR is sharing a programmer (anchored to U-M/ sustainability of projects)

11

External Funding

• $ 8.4M in MICDE backed-grants in 2015• $5M NSF DIBBS - McKee• $3.5M NSF MRI – Duraisamy

• $17.2M by 2016• $3.2M NIH U01- Figueroa• $2.5M NSF CRISP- El-Tawil

• $25.4M by 2017• $4.9M AFOSR/AFRL - Duraisamy• $2.7M TRI– Garikipati

• $37.4M by 2018• $1.8M NIH R01 – Aton & Zochowski• $7.2M NIH TCORS – Meza

$0.0

$10.0

$20.0

$30.0

$40.0

2015 2016 2017 2018

$8.7

$17.2

$25.4

$37.4

Cumulative (millions)

12

Catalyst Grants:Description

• High impact, innovative research projects in computational science

à potential to attract external funding

• Combine elements of mathematics, computer science, and cyberinfrastructure

13

Catalyst Grants:Research projects in any emerging area

14

Catalyst Grants:2017-2018 Successful Proposals

1. From Spiking Patterns to Memory Formation: Tools for Analysis and Modeling of Network-wide Cognitive Dynamics of the BrainPIs – Sara Aton (Developmental, Cellular and Molecular Biology)

and Michal Zochowski (Physics)

“” Collaborative research between the Zochowski and Aton labs has established a novel framework, built on more rapidly estimating network functional connectivity, to characterize the dynamics of memory encoding and storage…”

- $1.8M NIH R01 – Aton & Zochowski

15


2. Computational Energy SystemsPIs – Pascal Van Hentenryck, Eunshin Byon, Ruiwei Jiangand Jon Lee (IOE), and Johanna Mathieu (EECS)

3. Extreme, rare events in science & engineeringPI - Venkat Raman (Aerospace Engineering) in collaboration withRamanan Sankaran (Oak Ridge National Laboratory) and Jacqueline Chen (Sandia National Laboratory)

4. Integral equation methodsPI – Robert Krasny (Math)

16


1. Teaching autonomous machines to swim

PI Silas Alben (Math), Robert Deegan (Physics) and Alex

Gorodetsky (Aero)“” The research team will develop a computational and machine learning program to discover how to configure self-oscillating gels so that they undergo deformations that result in swimming. The long term goal is to develop a general framework for controlling autonomous soft machines..”

17


2. Exploring Quantum Embedding Methods for Quantum Computing Solution-adaptive, scale-aware climate models PIs Emanuel Gull, Physics; Dominika Zgid, Chemistry

3. Advancing the Computational Frontiers of Solution-Adaptive, Scale-Aware Climate ModelsPI Christiane Jablonowski, Climate and Space Sciences and Engineering; Hans Johansen, Lawrence Berkeley National Lab

4. Embedded Machine Learning Systems To Sense and Understand Pollinator BehaviorRobert Dick, EECS; Fernanda Valdovinos EEB, Complex Systems; Paul Glaum, EEB

18


5. Deep Learning for Phylogenetic InferencePIs Jianzhi Zhang, EEB; Yuanfang Guan, DCMB

6. Urban Flood Modeling at “Human Action” Scale: Harnessing the Power of Reduced-Order Approaches and Uncertainty QuantificationPIs Valeriy Ivanov, Nick Katopodes, CEE; Khachik Sargsyan, Sandia National Labs

7. Deciphering the meaning of human brain rhythms using novel algorithms and massive, rare datasetsPI Omar Ahmed, Psychology, BME

19

Catalyst Grants:Review Criteria

1. Work must be novel and not an incremental extension of existing work

2. Likelihood of success

3. Plan for specific external funding agencies or foundations to be approached as an outcome of the project

4. For full proposals: a plan to leverage the infrastructure of Advanced Research Computing (ARC) at the University of Michigan for external funding

20

Catalyst Grants:Budget and Justification

• Up to $100,000

• Post-doc and graduate students salary ~80%• Student Candidate rate only

• May include travel expenses, ARC resources

• No indirect cost or cost-sharing

• No PI/co-PI salary unless justified

21

Catalyst Grants:How to Apply

• Who• PI/co-PI must be an University of Michigan, Ann Arbor, faculty• Collaborative proposals are encouraged

• Pre-proposal • 2-pages project description – including references• Deadline to submit: February 18, 2019, 5:00 p.m. E.T.• Results expected by March 15, 2019

22

Catalyst Grants:How to Apply

• Full proposal – by invitation

• 6-pages project description• References• 2-page NSF or NIH style CV for PI and co-PIs• A detailed budget and budget justification

– template will be provided

• Deadline expected to be beginning of April• Results announced by end of April 2019• Project may begin on or after May 1, 2019

23

Catalyst Grants:Review Process

• 3-4 U-M members per review panel

• PIs should suggest 2-3 expert reviewers

• Review panels are provided a template with evaluation categories

• All proposals will get feedback, regardless of outcome

• Will aim to use same pool of reviewers for preliminary and full proposal evaluation

24

Catalyst Grants:FAQ

• Should the proposal include mathematics, computer science, and/or cyberinfrastructure collaborators? • Such collaborations are by no means required, but

encouraged. Cyberinfrastructure collaborations could come by engaging closely with ARC-TS or CSCAR.

• Must projects have new developments in computational methods, or are innovative applications of existing computational methods also acceptable?• Innovative applications of existing computational methods are

also sought.

• Can a proposal have a single PI?• Yes.

25

Catalyst Grants:FAQ

• Should PIs/co-PIs be MICDE affiliated faculty?• No, affiliation to MICDE is not required.

• Will other research areas not listed be considered?• Yes! Every research project with a relevant computational

component will be considered. However generic data-science are not suitable.

• If my project is chosen, is there a deadline for the start of the project?• All projects should start by September 1, 2019 to give PIs a

chance to recruit students/post-docs.

26

Catalyst Grants:FAQ

• Does the $100,000 have to be used up in 1 year, or could it be extended?• Extensions will be considered but PIs are strongly

encouraged to have a plan to use all the money in 1 year.

• Can any funds be used for lab analyses?• Data-analyses is ok, but at this time in-house experiments

cannot be funded through this grant.

• Does MICDE want software to be an output of the project?• No, software is not expected as an output of the project, but

it could be a component of the project.

27

Catalyst Grants:FAQ

• What are the deliverables?We will require teams to meet at the end of the project to share results, and discuss what worked/didn’t work in the process.NEW – present at MICDE Symposium in spring.

More Questions?

28

Catalyst Grants:More information

For more information please visit micde.umich.edu/grants/catalyst-grants/

29

Catalyst Grants:More information

Or contact:

Krishna Garikipati(ME) – [email protected]

Karthik Duraisamy(Aero)– Associate [email protected]

Annette Ostling(EEB) – Associate [email protected]

Siqian Shen(IOE)– Associate [email protected]

Mariana Carrasco-Teja– Assistant [email protected]

Thank you

Documents

The Michigan Institute for Computational Discovery ... · Graduate Certificate in CDE Graduate Certificate in Computational Neuroscience Ph.D. in Scientific Computing. 4 What is MICDE?