Regional CI
Click Here to Register
for the 21024 FLR
Annual Member Meeting
The objective of this “Regional Data Science Cyberinfrastructure” (Regional CI) partnership initiative is three-fold. Primary objective is to identify and catalog current data science cyberinfrastructure capabilities across South Florida. The second objective is to help amplify research and training capabilities in order to bolster harmonized data science education, and regional collaborations in science and engineering. The ternary objective is to help develop sustainable programs within the institutions and harness shared objectives across South Florida and beyond and ensure that every member of our academic community is empowered for the future.
The National Science Foundation’s (NSF) computational ecosystem of the 21st century identified that “Machine Learning and big data analytics are now an essential part of the scientific discovery process—complementary to, and increasingly integrated with, computational modeling. Simulation and modeling, measurement and instrumentation, and data analysis are deeply interdependent,” and disciplinary and cross-disciplinary collaborations require converged computing capacity for stronger engagement.
South Florida (within 50 miles of Miami) is the home for diverse research and educational capacity representing:
- Historically Black Colleges and Universities (HBCUs),
- Hispanic Serving Institutions (HSIs),
- Asian American, Native American and Pacific Islander -Serving Institutions (AANAPISIs), and
- Primarily Undergraduate Institutions (PUIs).
The main purpose of this regional data science CI effort is to bring together these institutions as closely as possible in terms of institutional and multi-institutional collaborations to support and strengthen academic programs, funding opportunities and supporting ways to build, enable, or strengthen sustainable campus CI capabilities.
Partnering Institutions
For this NSF funded Regional CI partnership development process, the University of Miami Frost Institute for Data Science and Computing collaborates with the following institutions and offering a series of workshops: Bethune-Cookman University (BCU), Florida Atlantic University (FAU), Florida International University (FIU), Florida Memorial University (FMU), Miami Dade College (MDC), and Nova Southeastern University (NSU).
Collaborators
Click the link above for a list of Core Participants, Extended Collaborators, and the External Advisory Committee.
Workshops
The planning grant involves hosting three one-day workshops. They are Collaborative Data Science and Education, identifying and evaluating Science Drivers, and Writing a CI Plan. The types of data that will be generated during these workshops include:
- Educational Materials (such as lecture notes, slide decks, tutorials, hands-on exercises) and multimedia (demonstration videos and recorded lectures)
- Software such as open-source community software and other application specific codes
- Datasets used for application specific testing and evaluation on computing resources
- Administrative data such as workshop survey results
- Metadata descriptions of the educational materials, software, and datasets
Collaborative Data Science and Education WorkshopThe Data Science and Education (DSE) workshop focuses on gathering insights from stakeholders on ways to organize existing data and programming tools with the goal of creating a repository of these tools for promoting research and education. Specifically it seeks collaborative solutions to two known problems in data science education:
The workshop draws from experience with the Software Carpentry (SWC) Foundation’s approach to teaching data science and research computing, and local institutional experience in leading training for AI and machine learning (ML) methodologies to solve problems in science and engineering. The outcome of the DSE workshop will be prioritized lists of needs for research training and skills development. The lists will provide the foundation for a repository of data science curriculum, tools, and datasets to support and promote collaborative data science research and education across South Florida. |
Science Drivers WorkshopThe one-day Science Drivers workshop aims to identify Science and Engineering applications that rely on regional cyber-infrastructure. Using the CC* regional computing template as a guide, efforts will be made to capture application requirements in terms of current and required computing and software requirements (both open source and commercial), common software tools used, and collaborators—especially from small and underrepresented institutions. This information will be used to develop a list of domain areas and evaluate their scope for interdisciplinary and multi-campus collaborations toward creating a sustainable, regional, data science cyberinfrastructure. This workshop will be structured to highlight example applications to train participants on submitting science drivers for evaluation. Research facilitation services will be provided during to help application researchers to evaluate their current and future computing needs, wherever feasible, using testbed resources. The research facilitators will use the insights to foster multi campus collaborations wherever such an effort advances application scope and impact. We will also showcase existing data science applications that are of interest to small and underrepresented institutions. |
Writing a CI Plan WorkshopThe workshop on writing a CI plan aims to educate, train, and engage institutional IT leaders and other administrators in decision making roles from small, minority-led, underrepresented, or under resourced institutions on the role, scope, and impact of a campus CI plan for enabling and supporting technologies for research and training. Existing CI plans and guidance from external advisors will be used as examples for this workshop. As a part of this workshop, Florida LambdaRail provides an overview on networking connectivity in concert with Science Drivers. They will also share experience working with IT leadership in evaluating current and developing needs. Speakers from the Minority Serving Cyberinfrastructure Consortium (MS-CC) and Sunshine State Education and Research Computing Alliance (SSERCA) will share their insights on the role of CI plan in streamlining campus IT investments and strengthening campus goals in terms of Data Science Education and Science Drivers. The workshop makes accessible new streams of funding opportunities to small, minority-led, underrepresented, or under resourced institutions. These include CC* grants for campus networking, storage, and computing infrastructure, and opportunities to leverage existing CC* programs such as MS-CC, especially for HBCUs. As a secondary outcome this workshop also assists in better understanding the regional CI plan that conveys the future vision for sustainable regional computing for S&E. |
Computing Environment
FAU, FIU, and UM will provide access to computing resources to test, and benchmark identified applications during this planning grant. Provided below list of resources along with their specifications and access methods for each of the listed resource from the collaborating institutions.
FIU RAPTOR. Raptor.FIU.edu. Up to 20% of RAPTOR (Reconfigurable Advanced Platform for Transdisciplinary Open Research) – a CC* resource – is available for this project. Co-PI Ibarra is the point of contact for resource and user training needs. All access requests for RAPTOR can be sent to HPCadmin@fiu.edu. RAPTOR provides four high-performance compute modalities to support transdisciplinary research.
FAU provides up to 1.8 million CPU-core hours through FAU’s High-performance computing environment. For more information, please visit hpc.fau.edu. Up to 300 GB of storage will also be available for application migration and support during the planning grant. Moreover, FAU’s
OwlCloud owlcloud.hpc.fau.edu offers to host any web sites or applicable web-based application or research Kubernetes Cluster. These resources will be leveraged for establishing web repositories and online data science training opportunities. Rhian Resnick, Director of Research Computing at FAU will be the point of contact for requesting access to the resources.
UM Four Intel Skylake nodes each with dual-socket Gold 5218 CPU running at 2.30GHz. Each node has a total of 32 CPU-cores and 131GB RAM. Up to 10TB network storage is available. Additionally, one of the 96 Triton idsc.miami.edu/triton nodes will be made available for this project as the need arises. Each Triton node is an IBM Power System AC922 architecture and equipped with two NVIDIA Tesla V100 GPUs and engineered to be “the most powerful training platform”. Triton utilizes the Power9 architecture which specializes in data intensive workloads.
Computing Facilitation
FIU, FAU, and UM will collaborate in providing research consultation services to application researchers. These activities include 1on1 consultations, matching application codes and computing resources as much as possible, providing access and user training on computing resources (see Computing Environment), and installation, testing, and benchmarking the application codes. This effort will assist in terms of:
- Assessing the current and future region-wide research infrastructure and capacity needs
- Limitations of computing and compatibility issues with respect to the software
- Identification of additional infrastructure for data staging and data movement, and
- Skilled workforce development opportunities and human-technology gaps.
Florida LambdaRail (FLR) Annual Member Meeting
Click for AGENDA | REGISTRATION Information
The FLR Annual Member Meeting takes place May 7-9, 2024 at Nova Southeastern University, Fort Lauderdale, Florida.