2024 NSF HDR Ecosystem Conference Harvesting the Data Revolution

9 Sept 2024, 09:00 → 12 Sept 2024, 22:00 US/Central

Illinois Conference Center (University of Illinois at Urbana-Champaign)

Mark Neubauer (Univ. Illinois at Urbana Champaign (US))

2024 NSF HDR Ecosystem Conference

NSF’s Harnessing the Data Revolution (HDR) Big Idea is a national-scale activity founded in 2016 to enable new modes of data-driven discovery that address fundamental questions at the frontiers of science and engineering.  By engaging NSF's research community in the pursuit of fundamental research in data science and engineering, the ecosystem of HDR Institutes, Data Science Corps (DSC), and Transdisciplinary Research in Principles of Data Science (TRIPODS) strive to provide a cohesive, federated, national-scale approach to research data infrastructure, and the development of a 21^st-century data-capable workforce.

The 3^rd annual conference of the NSF HDR Ecosystem will be held on the campus of the University of Illinois Urbana-Champaign from the 9^th to the 12^th of September 2024.

The main goals workshop are:

• Build and strengthen productive, inclusive, and positive relationships within the HDR ecosystem and the broader data-intensive research community.

• Provide a forum to share accomplishments, goals and plans of HDR ecosystem entities.

• Assess current and planned cross-cutting activities that address shared challenges and seek new opportunities to sustain and grow the HDR ecosystem and advance data-intensive research

• Assess the landscape of institutes, projects and activities within data science, AI, industry and infrastructure to advance a vision of how the HDR ecosystem is an integral part of a broader coherent ecosystem of AI and data-intensive research.

• Formally launch the HDR Machine Learning Challenge.

Announcements

Monday 9 September

1 Registration

The registration desk is in the Illinois Conference Center atrium, right in front of you when you enter through the main entrance via the parking lot.

2 Welcome Reception

Tuesday 10 September

3 Welcome and Introduction

Speaker: Mark Neubauer (Univ. Illinois at Urbana Champaign (US))

Welcome and Introduction

4 Welcome from the Dean of the College of Engineering

Speaker: Rashid Bashir (University of Illinois at Urbana-Champaign)

Dean-CS DS AI Q Deck.pdf

Dean-CS DS AI Q Deck.pptx

5 NSF Welcome and HDR Overview

Speaker: Amy Walton (National Science Foundation)

HDR - NSF Welcome - 10Sep2024.pdf

HDR - NSF Welcome - 10Sep2024.pptx

10:00 Coffee Break

HDR Institute Overview, Activities and Accomplishments

6 A3D3

Speaker: Shih-Chieh Hsu (University of Washington Seattle (US))

gslide

7 ID4

Speaker: Eric Toberer (Colorado School of Mines)

ID4 Fall 2024 HDR Ecosystem 15 minutes.pptx

8 Imageomics

Speaker: Tanya Berger-Wolf (The Ohio State University)

24-09 Imageomics HDR All Hands.pdf

24-09 Imageomics HDR All Hands.pptx

9 iHARP

Speaker: Vandana Janeja (UMBC)

iHARP OverviewJaneja.pptx

10 I-GUIDE

Speaker: Anand Padmanabhan (University of Illinois at Urbana-Champaign)

I-GUIDE HDR Conferance.pptx

11 NSF Panel

Moderator: Tanya Berger-Wolf (The Ohio State University)

A panel of the NSF program directors bringing the NSF perspective of the history and the future of the data revolution, particularly in the context of AI revolution.

Panelists primarily assembled from Cognizant Program Officers of NSF 21-519 HDR Institute and NSF24-560 DSC:

• Amy Walton Deputy Director for the Office of Advanced Cyberinfrastructure (OAC) at the National Science Foundation
• Sylvia Spengler, Program Director, Division of Information and Intelligent System (IIS) within the Computer and Information Science and Engineering (CISE) Directorate
• Chaitanya K. Baru, Senior Advisor, Directorate for Technology, Innovation and Partnerships (TIP)
• Raleigh Martin Program Director, Directorate for Geosciences (GEO) Division of Earth Sciences (EAR) Integrated Activities Section
• Cheryl L. Eavey Directorate for Social, Behavioral and Economic Sciences (SBE) Division of Social and Economic Sciences (SES) Methodology, Measurement, and Statistics (MMS)

12:30 Working Lunch: Networking

12 Keynote talk: Knowledge-Guided Machine Learning: A New Framework for Accelerating Scientific Discovery and Addressing Global Environmental Challenges

Climate change, loss of bio-diversity, food/water/energy security for the growing population of the world are some of the greatest environmental challenges that are facing the humanity. These challenges have been traditionally studied by science and engineering communities via process-guided models that are grounded in scientific theories. Motivated by phenomenal success of Machine Learning (ML) in advancing areas such as computer vision and language modeling, there is a growing excitement in the scientific communities to harness the power of machine learning to address these societal chal-lenges. In particular, massive amount of data about Earth and its environment is now continuously be-ing generated by a large number of Earth observing satellites, in-situ sensors as well as physics-based models. These information-rich datasets in conjunction with recent ML advances offer huge potential for understanding how the Earth's climate and ecosystem have been changing, how they are being impacted by humans actions, and for devising policies to manage them in a sustainable fashion. However, capturing this potential is contingent on a paradigm shift in data-intensive scientific discovery since the “black box” ML models often fail to generalize to scenarios not seen in the data used for training and produce results that are not consistent with scientific understanding of the phenomena.

This talk presents an overview of a new generation of machine learning algorithms, where scientific knowledge is deeply integrated in the design and training of machine learning models to accelerate scientific discovery. These knowledge-guided machine learning (KGML) techniques are fundamental-ly more powerful than standard machine learning approaches, and are particularly relevant for scien-tific and engineering problems that are traditionally addressed via process-guided (also called mecha-nistic or first principle-based) models, but whose solutions are hampered by incomplete or inaccurate knowledge of physics or underlying processes. While this talk will illustrate the potential of the KGML paradigm in the context of environmental problems (e.g., Ecology, Hydrology, Agronomy, climate sci-ence), the paradigm has the potential to greatly advance the pace of discovery in any discipline where mechanistic models are used.

Speaker: Vipin Kumar (University of Minnesota)

KGML_HDR Ecosystem202024 Keynote v2.pptx

Unconference "Elevator" Pitches

Open Space.pptx

15:30 Coffee Break

Transdisciplinary Research: Continuous ML and Human-in-the-Loop Decision Making

Convener: Sharad Sharma (University of North Texas)

Intro slides

Transdisciplinary Research (Continuous ML and Human-in-the-Loop Decision Making)

Transdisciplinary Research: Knowledge-Guided ML / Physics-Informed Neural Networks

Transdisciplinary Research: LLMs / Foundation Models for Research Minutes

Transdisciplinary Research: Responsible AI / Ethical AI Minutes

Unconference #1

Transdisciplinary Research: Knowledge-Guided ML / Physics-Informed Neural Networks

Conveners: Jianwu Wang (University of Maryland, Baltimore County), Zhijian Liu (University of California San Diego)

Intro slides

Transdisciplinary Research (Continuous ML and Human-in-the-Loop Decision Making)

Transdisciplinary Research: Knowledge-Guided ML / Physics-Informed Neural Networks

Transdisciplinary Research: LLMs / Foundation Models for Research Minutes

Transdisciplinary Research: Responsible AI / Ethical AI Minutes

Unconference #1

Transdisciplinary Research: LLMs / Foundation Models for Research

Conveners: Jiawei Han (University of Illinois at Urbana-Champaign), Pan Li (Georgia Tech)

Intro slides

Transdisciplinary Research (Continuous ML and Human-in-the-Loop Decision Making)

Transdisciplinary Research: Knowledge-Guided ML / Physics-Informed Neural Networks

Transdisciplinary Research: LLMs / Foundation Models for Research Minutes

Transdisciplinary Research: Responsible AI / Ethical AI Minutes

Unconference #1

Transdisciplinary Research: Responsible AI / Ethical AI

Conveners: Peter Darch (University of Illinois at Urbana-Champaign), Savannah Thais

Intro slides

Transdisciplinary Research (Continuous ML and Human-in-the-Loop Decision Making)

Transdisciplinary Research: Knowledge-Guided ML / Physics-Informed Neural Networks

Transdisciplinary Research: LLMs / Foundation Models for Research Minutes

Transdisciplinary Research: Responsible AI / Ethical AI Minutes

Unconference #1

Unconference Determined: Fair and Open Science

FAIR and Open Science

Lightning Talks for Posters

gslide

Poster Session

13 Imageomics: FAIR ML Products for Biological Knowledge Discovery

Speaker: Elizabeth Campolongo

HDR_Conference_Poster_2024.pdf

14 Incorporating phenotypic similarity into trait description embeddings​

Speaker: Soumyashree Kar

SKAR_HDRConf_Poster_Sep2024.pdf

15 Latent Space Phenotyping for Measuring Complex Evolutionary Traits

Speaker: Caleb Charpentier

calebcharpentier.pdf

16 Hierarchical Conditioning of Diffusion Models Using Tree-of-Life for Studying Species Evolution

Speaker: Mridul Khurana

Phylo-Diffusion.pdf

17 What Do You See in Common? Learning Hierarchical Prototypes over Tree-of-Life to Discover Evolutionary Traits

Speaker: Harish Babu Manogaran

HComP-Net_Poster.pdf

18 BioCLIP: A Vision Foundation Model for the Tree of Life

Speaker: Sam Stevens

19 Education and Outreach in Imageomics: Engaging Communities to Advance Science

Speaker: Diane Boghrat

Education and outreach in imageomics_HDR Conference.pdf

20 Using Deep Learning to Quantify Phenotypic Similarities in Mimic Butterfly Species using Human, Bird, and Butterfly Acuities

Speaker: Michelle Ramirez

21 Dynamic Network Classification

Speaker: Namrata Banerji

DynamicNetworkClassificationPoster.pdf

22 Tulane Center for Community-Engaged Artificial Intelligence

Speaker: Aron Culotta

tulane-ceai-poster.pdf

23 Understanding of impact of training size on animal re-identification

Speaker: Ekaterina Nepovinnykh

Poster_Imageomics_2024_Training_Size.pdf

24 Practical Leadership for Team Science: Experiences from the Imageomics Institute

Speaker: Diane Boghrat

Practical Leadership for Team Science - Experiences from the Imageomics Institute_HDR Poster.pdf

25 VLM4Bio: A Benchmark Dataset to Evaluate Pretrained Vision-Language Models for Trait Discovery from Biological Images

Speaker: Anuj Karpatne

VLM4Bio Poster.pdf

26 What Do You See in Common? Learning Hierarchical Prototypes over Tree-of-Life to Discover Evolutionary Traits

Speaker: Anuj Karpatne

HComP-Net_Poster.pdf

27 National Data Mine Network

Speaker: Mark Daniel Ward

NDMN_Poster_2024-25.pdf

28 Fish-Vista: A Multi-Purpose Dataset for Understanding & Identification of Traits from Images

Speaker: Kazi Sajeed Mehrab

Fish_Vista_HDR_2024.pdf

29 Low-cost Efficient Wireless Intelligent Sensors (LEWIS) for Engineering Education: Native American Knowledge for Data Science Education

Speaker: Fernando Moreu

moreu_UNM2.pdf

30 Facilitating Knowledge Sharing and Discovery: Search Functionality and API Design for the I-GUIDE

Speaker: Yunfan Kang

IGUIDE platform Search Spatial LLM.pdf

31 CLV: A Novel Framework for Enhanced Anomaly Detection and Attribution in Multivariate Time Series Data

Speaker: Tolulope Ale

Cluster_VAE_Poster.pdf

32 Battling Misinformation through Interdisciplinary Collaboration

Speaker: Zahra Khanjani

HDR2024Khanjani.pdf

33 CMAD: Advancing Understanding of Anomalous Melt Events over the Antarctic Sea Ice

Speaker: Maloy Kumar Devnath

CMAD.pdf

34 HDR DSC: The Metropolitan Chicago Data-science Corps (MCDC)

Speaker: Lizhen Shi

MCDC_Poster.pdf

35 Neural Network Efficiency Evaluation on the AMD Versal AI Engine

Speaker: Yilin Shen

AIE_HDRPoster.pdf

36 BaboonLand Dataset: Tracking Primates in the Wild and Automating Behaviour Recognition from Drone Videos

Speaker: Maksim Kholiavchenko

BaboonLand.pdf

37 Assessing Annotation AccAssessing Accuracy in Ice Sheets Using Quantitative Metrics

Speaker: Bayu Tama

Assessing Annotation.pdf

38 Genotype to Phenotype Mapping via Deep Learning

Speaker: David Carlyn

Genotype_to_Phenotype_48x36.pptx.pdf

39 Variance Analysis of Brightness Temperature using High-resolution DYAMOND simulations and CRTM in Digital Twin Systems

Speaker: Chhaya Kulkarni

Variance_Analysis_HDR_Conference_Poster.pdf

40 Predicting Sea ice extent over Antarctica using Patch CNN

Speaker: Sai Vikas Amaraneni

iharp_all_hands_alaska.pptx

41 HDR DSC: Collaborative Research: Transforming Data Science Education through a Portable and Sustainable Anthropocentric Data Analytics for Community Enrichment (ADACE) Program

Speaker: Yu Liang

UTC HDR DSC ADACE poster-Sept2024.pdf

42 Physics-Informed Sea Ice Thickness Prediction

Speaker: Akila Sampath

poster_NSF-HDR_AkilaSampath.pptx.pdf

43 Probabilistic Prediction of Material Stability: Integrating Convex Hulls into Active Learning

Speaker: Andrew Novick

Novick_Hull_Poster.pdf

44 Cyberinfrastructure for Scientific Data Preservation and Image Similarity Search

Speaker: Joshua Agar

Agar_Poster.pdf

45 Public Lecture: Making AI Safe, Effective, and Trustworthy

The data revolution has been replaced by the AI revolution. It seems that every day, we hear about a new area of human endeavor that has been conquered by AI. There are AI lawyers, AI doctors, AI artists, poets, and mathematicians. AI will predict what jobs we get, what medicines we should be treated with, and even how we should be educated.

If that's true, then AI policy should consist of getting the government out of the way and letting innovation bloom. Indeed, that is what some advocate. But in fact, sound AI policy rests on the same principles that sound science rests on: openness, transparency, and evidence. In order to ensure that all of us can benefit from innovations in AI, the best ideas in AI policy emphasize openness in research, transparency and accountability in the claims made by those seeking to deploy it, and above all a requirement that claims -- bold claims at that -- of efficacy are backed up by evidence that can be independently evaluated.

In this talk, I'll talk about how AI policy is trying to harness the AI revolution so that all of us, including those who have been traditionally left behind by tech innovation, can lead lives enriched rather than controlled by technology.

Speaker: Suresh Venkatasubramanian

hdr.pdf

Wednesday 11 September

AI & Data Infrastructure

AI & Data Infrastructure (Challenges and Opportunities for HDR Institutes and NAIRR Integration)

46 NSF OAC and the National Artificial Intelligence Research Resource (NAIRR) Pilot

Speaker: Katerina Antypas (National Science Foundation)

2024-09-11-HDR-Antypas.pptx

47 Frontier AI for Science Security and Technology (FASST) initiative

Speaker: Franck Cappello (Argonne National Laboratory)

FASST-UIUC-Sept-2024-for-Sharing.pptx

10:15 Coffee Break

48 Knowledge Transfer in AI Workload Acceleration: from Academia to Industry

At UIUC, there is strong ongoing collaboration between academia and industry, reflected in large industry-sponsored centers and institutes such as the IBM-Illinois Discovery Accelerator Institute (IIDAI) and the Center for Networked Intelligent Components and Environments (C-NICE), as well as in smaller and medium-sized partnerships like the AMD Center of Excellence and the Amazon-Illinois Center on AI for Interactive Conversational Experiences (AICE). The research activities in these centers are closely aligned with the strategic goals of industry partners, driving innovation, performance improvements, and market competitiveness across various high-tech sectors. Through close collaboration between UIUC students, faculty, and industry professionals, we aim to publish technical papers in top conferences and journals, while also transferring valuable knowledge back to industry. Additionally, many of these projects are open-source, with code freely available to benefit both the research community and industry. In this talk, Prof. Chen will highlight these topics, with a focus on AI workload acceleration.

Speaker: Deming Chen (University of Illinois at Urbana-Champaign)

HDR-DChen-9-24.pptx

Panel: The Role of AI, Computing and Data Infrastructure in Community Engagement

A panel conversation on the future of AI in relation to technology, computing, education, and inclusion, including experts across interdisciplinary domains spanning academia and industry.

Panelists:

• George Percivall GeoRoundtable: Spatial Web Foundation; IEEE Standards
• John Fonner Director of Special Programs, The Texas Advanced Computing Center
• Ashley Page Atkins Chief of Staff, San Diego Supercomputer Center
• Brett Bode Assistant Director, National Center for Supercomputing Applications
• Nhan Tran, Director of Real-Time Processing Systems Division, Fermi National Accelerator Laboratory

Convener: Anand Padmanabhan (University of Illinois at Urbana-Champaign)

Panel questions

Machine Learning Challenges

Notes

49 Machine Learning Challenges Overview

Speaker: Philip Coleman Harris (Massachusetts Inst. of Technology (US))

HDR ML Challenge (Link)

50 Detecting Novel Astrophysical Phenomena with Gravitational Waves (A3D3)

Speaker: Philip Coleman Harris (Massachusetts Inst. of Technology (US))

A3D3 ML Challenge Link

51 Detecting Anomalous Climate Phenomena (iHARP)

Speaker: Subhankar Ghosh (University of Minnesota)

Subhankar_HDRChallenge_iHARP_SeaLevelAnomalyDetection.pptx

52 Anomaly Detection: Hybrid Butterflies (Imageomics)

Speaker: Elizabeth Campolongo (The Ohio State University)

talk

12:30 Lunch

53 Pre-tenure faculty & postdoc mentoring lunch

54 Communicating Science: Effective Delivery To Any Audience

Abstract:
Effective science communication is crucial for bridging the gap between complex scientific concepts and diverse audiences. In this talk, I will explore strategies to convey scientific ideas clearly and engagingly, ensuring accessibility for individuals with varying levels of background knowledge. We will discuss the importance of storytelling in science communication, as well as techniques for simplifying technical jargon without compromising the integrity of the information. This talk aims to empower scientists, educators, and communicators to effectively share their work with broader, more diverse audiences, fostering a greater public understanding and appreciation of science.

Bio:
Sara Ayman Metwalli has a Ph.D. in Quantum Computing at Keio University, Japan, where she focused on developing and optimizing quantum algorithms and debugging tools for Noisy Intermediate-Scale Quantum (NISQ) devices. Sara has made significant contributions to quantum information science, particularly in the areas of quantum error correction and fault-tolerance. Her work has been published in leading journals and presented at international conferences, highlighting her role as an emerging leader in the quantum computing field.

In addition to her research, Sara has a strong commitment to STEM education and outreach. She has extensive experience as an educator, teaching programming and quantum computing to a wide range of students, from K-12 to university graduates. Sara is passionate about increasing diversity in STEM and has actively worked to create inclusive educational environments that empower underrepresented groups to pursue careers in science and technology."

Speaker: Sara Metwalli (Argonne National Laboratory)

Science Communication and Presenting_Sara.pptx

55 Promoting Educational and Outreach Activities throughout the HDR Institutes via a Centralized Repository

During the ideation expo at the previous HDR Ecosystem Conference in Colorado (Oct 2023), the creation of an HDR-wide educational and outreach materials repository was proposed, with a committee of cross-institutional members formed to build this repository. The long-term vision for the repository is to promote and facilitate data fluency in domain specific contexts for communities ranging from K-12th grade classrooms to the general public. In this talk, I will describe our efforts to design and implement a centralized repository to host educational and outreach materials collected throughout the HDR institutes. I will also report on a recent workshop (with cross-institutional participation) on the development of data fluency learning modules targeting 6-12th grade STEM classrooms, which we have included in the repository. Finally, I will describe our plans moving forward with the eventual goal of public promotion and release.

Speaker: Alex Pak (Colorado School of Mines)

AJP_HDR_Edu_Repo_Oct2024.pptx

Panel: Workforce Development and Education from Data Science Corps

This panel features representatives from five NSF-funded Data Science Corps (DSC) projects, each focused on harnessing data science to address real-world challenges while advancing education and workforce development. The panelists will share their experiences and insights on how their projects are making a significant impact in both academic and community settings.

Panelists:

• Francisco Iacobelli Associate Professor, Department of Computer Science, Loyola University Chicago
• Fernando Moreu Associate Professor, Department of Civil, Construction and Environmental Engineering, University of New Mexico
• Yu Liang Associate Professor, Department of Computer Science and Engineering, University of Tennessee, Chattanooga
• Amanda Phillips de Lucas Director, Baltimore Neighborhood Indicators Alliance, Jacob France Institute
• Mark Daniel Ward Professor of Statistics and (by courtesy) Mathematics, Purdue University

Convener: Eric Sokol (National Ecological Observatory Network (NEON), Battelle)

slides

15:30 Coffee Break

AI & Data Infrastructure: Challenges and Opportunities for HDR Institutes and NAIRR Integration

AI & Data Infrastructure (Challenges and Opportunities for HDR Institutes and NAIRR Integration)

HDR Student-Focused Programming

Convener: Diane Boghrat (The Ohio State University)

Minutes

Machine Learning Challenges: Future ML

Convener: Philip Coleman Harris (Massachusetts Inst. of Technology (US))

Notes

Unconference Determined: Model Validation, Interpretability, and Uncertainty Quantification

FAIR and Open Science

Workforce Development: Interdisciplinary Careers

Convener: Josephine Namayanja (UMBC)

Workforce Development Link

National Petascale Computing Facility Tour

NCPF_Tour.png

Conference Banquet

Fernando_Moreu_UNM.mp4

Thursday 12 September

56 Breakout session summary

breakoutUncertainty.pdf

Ethical AI

Foundational Models / LLM

Human in Loop AI

Knowledge-Guided ML / Physics-Informed Neural Networks

Workforce Development

10:00 Coffee Break

57 White Paper Planning

Speakers: Mark Neubauer (Univ. Illinois at Urbana Champaign (US)), Mark Stephen Neubauer (Univ. Illinois at Urbana-Champaign)

58 Conference Summary

Speaker: Philip Coleman Harris (Massachusetts Inst. of Technology (US))

PCH_NSFHDR_Summary_12_9.pdf

59 Closeout & Advertisement of 4th HDR Ecosystem Conference

Speakers: Mark Neubauer (Univ. Illinois at Urbana Champaign (US)), Diane Boghrat (The Ohio State University)

Closeout

next

11:40 Grab Boxed Lunches

Report Writing: Breakout #1

Report Writing: Breakout #2

Report Writing: Breakout #3

Report Writing: Breakout #4

National Petascale Computing Facility Tour

NCPF_Tour.png