Data Science Updates is the University of Wisconsin-Madison's resource for news, training, events, and professional opportunities in data science, brought to you by the Data Science Institute, powered by American Family Insurance, and the Data Science Hub.
April 30, 2025
|
|
|
|
Join the Data Science Hub on May 19-22, from 9:00 a.m. - 12:30 p.m., at the Discovery Building for an introductory workshop on the basics of deep learning. Upon completion, you will be able to train your first neural network. We will focus on the Keras framework — an excellent choice for those who don't want to write 300 lines of code to fit a single neural network model. Participants should be comfortable coding in Python and understand machine learning fundamentals. Register by May 14th.
|
|
|
|
Join the Data Science Institute and the Department of Mathematics on May 16 for a special seminar with Dr. Talitha Washington, professor of mathematics at Howard University. Dr. Washington will share insights on advancing AI technology and policy responsible for ensuring these technologies promote innovation, accountability, and broad societal benefit. This seminar will be held from 2:00-3:15 p.m. in the Discovery Building Orchard View Room. Coffee and cookies will be served.
|
|
|
|
Jelena Diakonikolas, assistant professor of computer sciences, wants to strengthen the algorithms that power machine learning models. She recently received an NSF CAREER Award to develop new algorithms that perform well under uncertainty and change. Her CAREER award was inspired by a Data Science Institute symposium where researchers voiced challenges with building models that are robust in ambiguous and changing data environments, echoing broader issues with how machine learning models are trained and used today. Read her story in Faces of Data Science.
|
|
|
|
TODAY, April 30, 10:00 a.m. - 11:00 a.m.; Zoom. The Posit Team will walk through how {pins} can streamline asset sharing across teams and projects, especially when you're working with Databricks. You'll learn to publish models, datasets, and other assets to Databricks Volumes. The team will also teach you to integrate {pins} into R workflows and how to use Spark to accelerate predictions with a 'pinned' model.
|
|
|
|
May 2, 9:30 a.m. - 4:30 p.m.; 1360 Biotech. The workshop will provide a hands-on introduction to software and analysis pipelines for RNA-Seq. Participants will learn the computational process that takes the raw data through the high level analysis. We will focus on advanced analysis using a Linux command line environment to run open-source RNA-Seq software.
|
|
|
|
Have questions about anything data science-related? Come see the Data Science Hub facilitators at Coding Meetup on Tuesdays and Thursdays from 2:30-4:30 p.m. CT. To join Coding Meetup, join data-science-hubgroup.slack.com.
|
|
|
|
TODAY April 30, 12:30 p.m. - 1:30 p.m.; Orchard View Room, Discovery Building & Zoom. Professor Dr. Ben Grimmer, Johns Hopkins University, will describe the best gradient method for smooth convex optimization. Dr. Grimmer will present a “subgame perfect” method that is optimal against a worst-case problem instance and leverages all gradient information revealed at each step. Finally, he will address the problem of optimally selecting stepsizes for gradient descent.
For those who have not signed up to attend in-person, please refrain from taking pizza, as catering is arranged beforehand.
|
|
|
|
TODAY, April 30, 1:30 p.m. - 2:30 p.m.; 159 (Wisconsin Idea), Education Building. AI applications improve our quality of life; however AI also perpetuates bias and harms many populations. Children must be involved at an early age in learning about how such technology works from a human-centered and sociocritical lens. Dr. Golnaz Arastoopour Irgens, Vanderbilt University, will discuss her research on designing AI literacy activities with and for elementary and middle-school-aged children that integrate social, ethical, and ideological dimensions.
|
|
|
|
TODAY April 30, 4:00 p.m. - 5:00 p.m.; 133 Service Memorial Institute. BART is an effective nonparametric regression model that approximates unknown functions with a sum of binary regression trees. Dr. Sameer Deshpande, UW–Madison, will describe BART's extensions based on new priors for decision rules and leaf outputs that overcome limitations. Dr. Deshpande will use BART's extensions to show the effects of playing American-style tackle football in adolescence on adult-life health.
|
|
|
|
TODAY April 30, 6:00 p.m. - 7:00 p.m.; 1240 Computer Sciences. Undergraduate Projects Lab will share internship advice, resume tips, and their experiences finding internships. They will describe the current job market and strategies you can take to prepare yourself for the upcoming season. Pizza will be provided.
|
|
|
|
May 1, 8:00 a.m. - 1:00 p.m.; Union South & Zoom. Register for and join UWEBC at the Wisconsin Digital Symposium, a high-impact event for tech and business leaders navigating today’s fast-moving digital landscape. From AI-powered automation to product-led transformation, you’ll gain strategies that make a difference—plus insights from keynote speakers Dr. Michael Proksch, global AI expert and Chief Scientist at AccelerEd, and Cortney Thompson Rowan, EVP of Strategy & Design at Delve.
|
|
|
|
May 1, 11:00 a.m. - 11:50 a.m.; 901 Van Vleck Hall. Dr. Elisenda Grigsby, Boston College, will discuss how the parameter space for any fixed architecture of neural networks serves as a proxy during training for the associated class of functions. She will discuss two locally-applicable complexity measures for Rectifier (ReLU) network classes and what her team knows about their relationship.
|
|
|
|
May 1, 4:15 p.m. - 5:15 p.m.; 1163 Mechanical Engineering. Dr. Lejo Flores, Boise State University, will discuss how cloud microphysics schemes simulate hydrometeor evolution and control precipitation and radiation predictions. Dr. Flores will present research evaluating how the choice of cloud microphysics scheme influences predictions of mountain snowpack evolution and report on implications for hydrologic and atmospheric research.
|
|
|
|
Apply by May 1- The Data Science Hub is looking for volunteers to help during the MLM25 hackathon. The MLM25 will run from September 11 to December 11 and bring together teams of all skill levels to work on applied ML/AI challenges drawn from research, industry, and open-source communities. We need administrative help, project organizers, advisors, and presenters. Volunteers will meet 2-3 times per month, starting in May, leading up to the event.
|
|
|
|
May 2, 2:00 p.m. - 3:00 p.m.; 175 Science Hall. Join Dr. Gengchen Mai, Director of the Spatially Explicit Artificial Intelligence (SEAI) Lab at the University of Texas at Austin, for a presentation on recent works from the SEAI Lab about spatial representation learning, including various location encoding models, an SRL deep learning framework (TorchSpatial), and a geo-foundation model (GAIR).
|
|
|
|
May 2, 2:25 p.m. - 3:25 p.m.; 901 Van Vleck Hall. HJ reachability is a powerful tool for solving differential games with bounded inputs, offering safety and liveness guarantees. However, it suffers from the “curse of dimensionality.” Dr. Sylvia Herbert, UC San Diego, will explore two methods for overcoming this issue: conservative guarantees and probabilistic guarantees.
|
|
|
|
May 6, 12:00 p.m. - 1:00 p.m.; Orchard View, Discovery Building. The ML + X community invites Dr. Benjamin Lengerich, UW-Madison statistics professor, and Kaiser Pister, UW-Madison computer science lecturer, to present research, teaching, and business perspectives on generative AI. Register by May 2nd to guarantee your lunch ticket.
|
|
|
|
May 7, 9:00 a.m.- 11:00 a.m.; 1145 Discovery Building. Join the ML + X community for SQLDeps and RAG demonstrations. Dr. Ron Stewart, Morgridge Institute for Research, will explore whether we can evaluate PubMed abstracts using LLM analysis with RAG based on co-occurrence models to provide some level of support or refutation for a hypothesis more quickly.
|
|
|
|
May 8, 12:00 p.m. - 1:00 p.m.; 175 Science Hall. Dr. Yuhao Kang, Director of the GISense Lab at the University of Texas at Austin, will explore the impact of Generative AI on geospatial analytics. Dr. Kang will discuss how a Soundscape-to-Image model could translate and visualize human perceptions of visual and acoustic environments. He will also illustrate how generative AI, through data-style separation, can produce accurate and visually appealing maps that adhere to ethical standards in cartography.
|
|
|
|
May 16; UIC. This symposium unites leading experts and practitioners for a dynamic program of keynotes, talks, fireside chats, and panels. We will explore cutting-edge AI applications, responsible strategies, and workforce implications that ensure AI drives business success and societal benefits. There will be s essions on applied AI, interpretability, energy infrastructure, and responsible innovation.
|
|
|
|
May 30; Discovery Building. This retreat is required for CIBM trainees and Biomedical Data Science PhD students in year two and later. Hear from keynote speakers David Page, Duke University, and Sriraam Natarajan, University of Texas at Dallas. Showcase your research and network with our community by presenting a poster. Register for the retreat and submit poster abstracts by May 9th.
|
|
|
|
Register by May 31- Join the National Center for Quantitative Biology of Complex Systems for the annual North American Mass Spectrometry Summer School from July 21-24 at the Discovery Building. Students will experience an engaging program covering fundamentals of mass spectrometry and the latest in its application to the analyses of plants (NSF) and animals (NIH). There will be lectures and workshops for scientific and professional development.
|
|
|
|
June 2-6; Fluno Center & Zoom. HTC25 brings together researchers, campus, science collaborations, facilitators, administrators, government representatives, and professionals interested in high throughput computing to:
- Engage with the throughput computing community, including the OSG Consortium, Center for High Throughput Computing, HTCondor staff, PATh and Pelican teams, and others contributing to HTC
- Be inspired by presentations and conversations with community leaders and contributors sharing common interests
- Learn about HTC and new developments to advance your science, your collaboration, or your campus
|
|
|
|
June 20 - July 25, 1:00 p.m. - 4:00 p.m. The Sky's the Limit STEM Camp, hosted by the Nelson Institute Center for Climatic Research, broadens science opportunities for autistic youth in grades 5-12 with a medical diagnosis, self-diagnosis, or suspected diagnosis. The camp provides nature-based and interactive learning opportunities to build interest and appreciation for STEM (science, technology, engineering, and mathematics). Attendees, accompanied by their caregivers, will participate in science experiments and outdoor activities. Registration is limited to 20 participants.
|
|
|
|
|
|
|
DATA VISUALIZATION OF THE WEEK
|
|
|
|
In 2024, wind and solar reached a record 17% (757 TWh) of US electricity, overtaking coal for the first time, which dropped to a historic low of 15% (653 TWh). Just six years ago, in 2018, coal was three times larger than the combined total of wind and solar. Solar generation boomed in 2024, rising by 27%, while wind rose by 7% and coal fell by 3.3%. Since the peak of US coal power in 2007, wind and solar have overtaken coal in 24 states, with Illinois the latest to join the ranks in 2024, following Arizona, Colorado, Florida and Maryland in 2023.
|
|
|
|
Data Science Updates is a collaborative effort of the Data Science Institute and Data Science Hub. This newsletter was originally created by the Data Science Hub and published as Hub Updates.
Use our submission form to send us your news, events, opportunities and data visualizations for future issues.
|
|
|
|
|
