NASA is running a new round of ROSES for supporting open source software. If you would like to contribute to the Julia community, one of the best ways would be to secure grants to get more students working on the ecosystem!
SMD seeks proposals for the improvement and sustainment of existing high-value,
open-source tools, frameworks, and libraries that have significantly impacted the SMD
science community. We are seeking proposals that satisfy the following objectives:
• Open-source software tools, frameworks, and libraries that have significant
usage in the NASA science community, were developed following open and
collaborative practices, and are aligned with the scientific vision and data
strategic plan of SMD as described in Science 2020-2024: A Vision for Scientific
Excellence - 2023 Update and the SMD’s Strategy for Data Management and
Computing for Groundbreaking Science 2019-2024 documents .
• Proposals should look to improve the sustainability and utility of these tools,
frameworks, or libraries through improvements such as adding extensions,
documentation, infrastructure, security, and maintenance of the software.
For the purposes of this program element, the following classifications apply:
• Tools: Systems that support scientific processes and analysis. This can include
software packages, web- or cloud-based tools, or other digital services. This can
include tools related to publications, data, software, events, data analysis,
modeling, machine learning, or other scientific processes.
• Frameworks: Systems that incorporate a variety of inputs to enable scientific
processes. This can include data formats, collections of models, citizen science
infrastructure, or other digital services.
• Libraries: Generic software packages, often with a larger user base,
implementing well-known algorithms, providing statistical analysis, visualization,
or other services that are incorporated in other software or used on their own.
I’d be happy to chat with any academics (i.e. professors, staff, or postdocs who can have PI status) who are interested in getting involved but haven’t done this before.
A few of us in the Julia Astro community are considering it. We’re having a Julia Astro Zoom meeting at noon (EDT) on Thursday, April 25th. This solicitation may be part of the discussion. All are welcome to join us.
Note that NASA held a Zoom meeting about this solicitation on April 2nd.
Unfortunately I’m not based in US, and I’m not an academic either (just a research engineer) but I’ve been using JuliaGNSS a lot: JuliaGNSS · GitHub for:
low Earth orbit GNSS reflectometry, the idea is to analyze reflected GPS signals from the Earth surface to determine its properties, such as ocean winds and soil moisture
lunar PNT, trying to navigate using GPS signals on Moon orbit. We also use SPICE.jl to simulate the orbits
I wish we have more resources to write high-quality reference documentation and guided tutorials on how to use it and how to implement new algorithms
Is this US-specific? If so, are you aware of other similar grants that are accessible to folks outside of the US?
I’m starting to consider applying for funding myself, in order to boost the JuliaGraphs and JuliaDiff ecosystems.
It’s for US-based institutions, but you can get crafty. For example, with JuliaGraphs I know @jpfairbanks has been a maintainer for a long time, so you could possibly rope him into taking lead with you as Co-I but help with a lot of the technical writing and mentoring of the eventual student?
Is there someone from US, who would be interested in work on Metatheory.jl?
I am interested in learning a heuristic function to select rewriting rules. Ultimately, I would like to have something like GitHub - facebookresearch/neural-rewriter: Learning to Perform Local Rewriting for Combinatorial Optimization, but taking advantage of flexibility of metatheory, such that optimizing the heuristic for a given theory would be easy. Since I have an experience with learning for planning (we have a paper on Neurips this year on that) and for learning over relational domain, this problem is a perfect fit.
While I can contribute to the project my academic time paid from other sources, I am looking for support of my former student. Since I am in Prague, Czech Republic, I cannot apply. Thanks for interest in advance.
Note that to be successful, you want to focus on high-impact tools which you can showcase are part of NASA missions or potential NASA missions. Metatheory.jl is a bit too theoretical in that sense. A successful application for it would probably need letters from some folks at NASA describing how they are currently using it and what improvements would be the most helpful.
Instead, ROSES is really built for general-purpose tooling. Things like differential equation solvers, bindings of Julia to Python, standard file IO readers, automatic differentiation and deep learning libraries immediately come to mind.
I’d be interested in applying, though my Stanford postdoc is wrapping up in September and I cannot work on other projects during the postdoc. I’d probably be interested in doing any number of Julia things ranging from the DL stack, general autodiff, stats stuff, or any of our core infrastructural packages.
So, not going to be academic-ish without something else happening. Anyone have insight on what might work there? I’d rather stay and contribute to the Julia community if I can.
This program element supports two types of awards: Foundational Awards and Sustainment Awards. The former is for open-source tools, frameworks, and libraries that have a significant impact on two or more divisions of the SMD. They are for up to 5 years. The latter is for open-source tools, frameworks, and libraries that have significant impact in one or more divisions of SMD. They are for up to 3 years.
My thoughts are as follows.
For a Foundational Award, I’m thinking about the packages that are applicable to all wavelengths of astronomy. Astrometry, IERS, Time, Unitful, FITS, and time-series analysis packages come to mind. Basically, any astronomy related package that is fundamental to astronomy.
For a Sustainment Award, I’m thinking of spectral analysis. My current idea is a Julia implementation of Xspec, the spectral analysis tool for X-ray astronomy, that is based on the SpectralFitting.jl package out of Bristol. Although Xspec is continually maintained, it has several limitations in the way it does optimization and statistics. For example, it does not handle Poisson statistics well. The idea is to initially develop the package for X-ray spectroscopy with the long-term goal of generalizing it to other frequency bands.
In the case of international participation, those members would be listed as unpaid consultants. If the proposal is approved, then they may want to apply to their national agencies for funding.