[SciPy-User] Call for collaborators in NSF grant proposal for scientific software for modeling complex dynamical systems with Python

[Rob Clewley]

Dear colleagues,

I am planning to put in a grant application to the NSF’s Software
Infrastructure for Sustained Innovation (SI^2) program, due by July
18th. The solicitation encourages a multi-disciplinary team, and
although a Program Officer was supportive of my idea in itself, he
encouraged me to find one or two others from fields outside my own
(computational/mathematical neuroscience) who would be interested in
pushing the idea into other disciplines with a longer term view of
development. I understand that there isn’t much time, but maybe some
of you familiar with dynamical systems modeling, model optimization,
and complex systems, might be interested in discussing further with me
offline. I have a broad draft of a proposal ready, and your ideas to
expand on it further with a vision that resonates with mine could
still be easily incorporated in time for the deadline.

I wish to contribute towards a solution to the broad problem faced in
many areas of modern modeling: “We don’t know how to do data-driven
science.” Specifically, my interest is in developing better
computational tools to explore and diagnose hypothesized mechanisms in
high-dimensional nonlinear dynamical systems. For the most part, my
interest lies with models defined by (ordinary) differential
equations, rather than discrete mappings or automata. You may be using
standard simulators or specific technology for your field, and I am
not proposing to build a more efficient simulator for large-scale
models. I want to develop pioneering new tools for data-driven
modeling that interfaces with a simulator and traditional analysis
tools (maybe ones that are specific to your scientific field, but are
likely to include bifurcation analysis, qualitative geometric
analysis, fast-slow multi-scale reductions, statistical analysis). The
new tools are expected to introduce algorithms incorporating
qualitative reasoning and heuristics to help a user better perform
their difficult work. (My opinion is that computer-assisted approaches
explicitly involving a supervising expert user can make much more of a
short term impact than fully automated approaches, which I think are
currently too ambitious.)

What I am seeking from a collaborator (a faculty member at a US
research institution eligible to be a co-PI) is a resonant interest in
forward-looking questions about understanding and engineering
large-scale dynamic models based on a modular building-block approach
(i.e., both analysis and synthesis). You would be interested in
Python-based tools in a similar spirit to mine (see below), which
would be directly relevant to your personal scientific interests in a
different disciplinary area. E.g., this could be coming from areas
such as climate modeling, geophysics, biochemistry, genomics,
biomechanics, astrophysics, all of which share problems of managing
complexity in large, nonlinear models involving mixed levels of
representation and multiple scales. My proposal springboards off
several ideas already prototyped in my PyDSTool dynamical systems
software environment (http://pydstool.sourceforge.net), but the aims
of the proposal do _not_ need to be focused solely on development
within the PyDSTool framework itself.

I currently have an NSF award for a related project, not focused on
the software tools themselves, and some publications that have come
from it. I also have two students working on applications of the
software in this project. Some of the developments are fueling this
new grant proposal. You can read more about my research here:
http://www2.gsu.edu/~matrhc/. You can read more about the NSF
solicitation here: http://www.nsf.gov/pubs/2011/nsf11539/nsf11539.htm.
I can send you my proposal draft if you contact me with your interests
(rclewley AT gsu DOT edu) and I am available to talk on the phone or
skype next week.

Thanks for your attention, and I look forward to talking with you!
Regards,
Rob


-- 
Robert Clewley, Ph.D.
Assistant Professor
Neuroscience Institute and
Department of Mathematics and Statistics
Georgia State University
PO Box 5030
Atlanta, GA 30302, USA

tel: 404-413-6420 fax: 404-413-5446
http://www2.gsu.edu/~matrhc
http://neuroscience.gsu.edu/rclewley.html