|
|
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.
October 30, 2024
|
|
|
|
Seminar Spotlight: Machine Learning for Medical Imaging
Machine learning (ML) techniques hold unprecedented potential to solve challenging problems in medical imaging research. The Machine Learning for Medical Imaging (ML4MI) initiative fosters interdisciplinary collaboration between ML experts and medical imaging researchers at UW–Madison. ML4MI hosts monthly seminars describing technical developments in ML with potential biomedical applications, radiology problems that may benefit from ML approaches, radiology projects involving ML techniques, and pioneering ML applications in biomedical research. Learn more and view the seminar schedule at the ML4MI website.
|
|
|
|
Faces of Data Science: Damon Smith
|
|
|
|
Apply to be a DSI Affiliate
The Data Science Institute is accepting applications for our affiliates program through December 1. DSI Affiliates are UW-Madison faculty and academic staff who are actively engaged in foundational or use-inspired research relevant to data science and aligned with the DSI mission and vision. The institute emphasizes cross-disciplinary collaboration and approaches, and we welcome affiliates from departments and centers across campus. Visit the DSI website to learn more and apply.
|
AI for Music and the Humanities: Insights and Impacts
November 5, 12:00 p.m. - 1:00 p.m.; Wisconsin Institute for Discovery, Orchard room 3280 + Zoom.
In this month's ML+X forum, we explore two perspectives on how artificial intelligence is reshaping music and the humanities. Alan Ng presents his application of machine learning to identify Irish traditional dance tunes, demonstrating the power of AI (conformer models) to tackle unique challenges within cultural preservation and musicology. By adapting a state-of-the-art model for folk music, Ng has created a tool that can instantly recognize melodies, supporting musicians and scholars alike. Complementing this, media studies researcher Jeremy Morris delves into the broader societal impact of AI in the music industry, examining how emerging technologies intersect with issues of creativity, intellectual property, and the evolving role of human agency in a platform-driven music landscape. Together, these talks highlight both the practical applications and the broader implications of AI within the humanities, sparking conversation around the opportunities and ethical considerations at this interdisciplinary frontier. Register by 11/3 (lunch provided) to guarantee your lunch ticket and join the discussion!
|
|
|
|
Methods for Biological Data Workshop
November 7 and December 5, 1:00 p.m. - 2:30 p.m.; Russell Labs, Room 584 + Zoom. Do you have too much data or computation to run on your laptop or lab server? One campus resource to scale up your computing is CHTC – the Center for High Throughput Computing. In this workshop, we will briefly cover when CHTC is a good home for your computing workflows and then do a hands-on demonstration of how data analysis can be run in CHTC. Check out the workshop website for more information.
Please bring your laptop. You will need access to a Unix (Apple)/Linux bash command-line and basic shell computing knowledge. If anyone needs help getting set up, facilitators Emile and Claudia will be in Russell Labs 584 and the Zoom room at 15 minutes before the workshop starts.
|
|
|
|
|
Registration for Mini-Workshop Series
Registration is open for the Data Science Hub's Fall 2024 Mini-Workshop Series. The Mini-Workshop Series are one-to-two day workshops throughout the Fall covering a range of topics listed below. All workshops take place online from 9am-1pm. Register for any and all that you are interested in. Tickets close the Friday before each mini-workshop. To learn more and register, visit the mini-workshop event page.
We are also looking for a couple of ML-savvy folks interested in helping out with the November 20-21 workshop (during one or both days). Helpers will help troubleshoot bugs and answer questions from learners during the workshop. We typically try to have at least 1 helper for every 5 learners at the workshop. This ratio helps us make sure that no learner gets left behind during the workshop! Email ml-community-of-practice@g-groups.wisc.edu if you are interested.
|
| Topic |
Date & Time |
Location |
| Interactive Data Visualizations in Python and Streamlit |
November 6
9:00 a.m. - 1:00 p.m. |
Zoom |
| Intro to Machine Learning with Sklearn |
November 20-21
9:00 a.m. - 1:00 p.m. |
Zoom |
|
|
|
|
Posit Day 2: Intro to Shiny Apps – Python Focus
November 19, 1:00 p.m. - 3:00 p.m.; Zoom. This is the second installment of the Fall 2024 Posit Days. Ryan Johnson will talk to us about Shinny apps using Python and touch on concepts such as: User interface- Inputs & Outputs, Server Logic, Reactivity, and Layout and Style. This event will be online and open to everyone affiliated with UW–Madison. Learn more and register for Posit Day 2.
|
|
|
|
Have questions about anything data science-related? Come see the Data Science Hub facilitators at Coding Meetup on Thursdays from 2:30-4:30 p.m. CT. To join Coding Meetup, join data-science-hubgroup.slack.com
|
|
|
|
SILO: Recent Advances in Min-max Optimization: Convergence Guarantees and Practical Performance
TODAY October 30, 12:30 p.m. - 1:30 p.m.; Wisconsin Institute for Discovery, Orchard room 3280 + Zoom.
Join professor Nicolas Loizou from John Hopkins University as he discusses how min-max optimization plays a prominent role in game theory, statistics, economics, finance, and engineering. Stochastic Gradient Descent Ascent (SGDA) and Stochastic Extragradient methods (SEG) have recently gained significant interest from the machine learning community. SGDA and SEG rank among the most efficient algorithms for solving large-scale min-max optimization and variational inequality problems (VIP) that occur in various machine learning tasks.
However, SGDA and SEG require strong assumptions and leave many important questions unanswered. In this talk, Loizou will address these questions and provide novel convergence guarantees for several variants of SGDA and SEG, diving into the details of their efficient implementations.
|
|
|
|
Biostatistics and Medical Informatics Department Seminar with Hannah Wayment-Steele
November 1, 12:00 p.m. - 1:00 p.m.; Genetics-Biotechnology Center Building, Auditorium. The functions of biomolecules are often based in their ability to convert between multiple conformations. The next frontier lies in how well we can characterize, model, and predict protein dynamics. Professor Wayment-Steele will discuss a simple adaptation of AlphaFold to predict multiple conformations, development of large-scale benchmarks of dynamics from across multiple types of NMR experiments, and whether it may already be possible to predict the present of biologically-relevant motions. Visit the Biostatistics and Medical Informatics website and the Professor Wayment-Steele's seminar calendar listing to learn more.
|
|
|
|
Distinguished Lecture: Gradient Optimization Methods: Large Step-Size and the Edge of Stability
November 4, 4:00 p.m. - 5:00 p.m.; 1240 Computer Sciences + Zoom. Peter Bartlett, professor of statistics and computer science at UC Berkeley, and principal scientist at Google DeepMind, will discuss how optimization in deep learning relies on simple gradient descent algorithms. Although these methods are traditionally viewed as a time discretization of gradient flow, in practice, large step sizes - large enough to cause oscillation of the loss - exhibit performance advantages. This talk will review recent results on gradient descent with logistic loss with a step size large enough that the optimization trajectory is at the "edge of stability," and show the benefits of this initial oscillatory phase for linear functions and for two-layer networks. Based on joint work with Yuhang Cai, Michael Lindsey, Song Mei, Matus Telgarsky, Jingfeng Wu and Bin Yu. For more information, visit the Gradient Optimization Methods seminar calendar listing.
|
|
|
|
Saving Time with AI: Smart Ways to Initiate Projects and Lessen the Administrative Burden
November 6, 2:00 p.m. - 3:00 p.m.; Zoom. Project initiation can be an administrative burden for project managers. This session will offer some tips and tricks on how to use Artificial Intelligence (AI) to cut down on the administrative burden of project initiation. It will explore how AI can help you ask the right questions to properly initiate a project and avoid common pitfalls that lead to wasted time and resources. Visit the Office of Strategic Consulting's website and the Saving Time with AI seminar calendar listing to learn more.
|
|
|
|
Statistics Seminar: Statistical learning for model-agnostic searches for new physics at the Large Hadron Collider
November 6, 4:00 p.m. - 5:00 p.m.; 133 Service Memorial Institute. Join Mikael Kuusela, assistant professor of statistics and data science at Carnegie Mellon, as he gives an overview of his recent work on model-agnostic searches for new physics in high-dimensional feature spaces. Searches for new phenomena at the Large Hadron Collider at CERN usually boil down to performing a statistical hypothesis test for the presence of a new signal over a background of known physics. Due to the high dimensionality of the feature space, these tests are usually done with the help of machine learning classifiers. This method is well established with reliable samples from both the signal and background distributions. However, Kuusela's team developed powerful tests that make weak assumptions about the signal, the background or both. Throughout this talk, Kuusela will draw connections to high-dimensional two-sample testing, anomaly detection, transfer learning and simulation-based inference. For more information, visit the statistics seminar calendar listing.
|
|
|
|
David DeMets Lectures
November 7, 3:30 p.m. - 4:45 p.m.; HSLC room 1325; reception after in HSLC Atrium until 6:00 p.m.
November 8, 12:00 p.m. - 1:00 p.m.; Biotechnology Center Auditorium. The 8th annual David L. DeMets Lectures in Health and Quantitative Investigation Lectures will consist of one general presentation on November 7 and one technical presentation on November 8. These lectures acknowledge Dr. DeMets’ research accomplishments in forwarding the design, monitoring, analysis, and presentation of clinical trials.
The first lecture is a research/overview talk is accessible to a general biomedical/public health audience. Ideally, it will highlight in one or more ways critical and impactful contributions from quantitative methodological areas (e.g., biostatistics and/or biomedical informatics) to discoveries in biomedicine, and/or advancement of human/public health. The second lecture is a world-class research talk in biostatistics or biomedical informatics.
|
Computer Sciences Department Research Symposium
November 11, 9:00 a.m. - 4:00 p.m.; Computer Sciences Building. Come join the CS Department Research Symposium to learn about the wonderful research taking place around the department. Also listen in on some key notes from our faculty! For more information, visit the computer science calendar listing.
|
ML4MI Seminar: Diffusion Bridge Models for CT Image Restoration and Reconstruction
November 12, 11:00 a.m. - 12:00 p.m.; Zoom. Join Dr. Dufan Wu, assistant professor of radiology at Harvard Medical School, in discussing his team's recent work on developing diffusion bridge models for CT image restoration and reconstruction. Inspired by DDIM, his team proposed a non-Markovian chain-based diffusion bridge model for image restoration, “Implicit Image-to-Image Schrodinger Bridge (I3SB)”, achieving a 2 to 4 times acceleration compared to vanilla bridge models. To improve data fidelity and reduce hallucination, his team proposed a novel method to incorporate data constraint into the reverse sampling of diffusion bridge models in a cost-function framework. Applications including denoising, sparse-view reconstruction, and motion correction were used to demonstrate the efficacy of the methods.
|
|
|
|
|
|
ML+Coffee: How Can I Apply Machine Learning to My Data?
November 6, 9:00 a.m. - 11:00 a.m.; Rm. 1145, Discovery Building. As part of the ML+X community's monthly ML+Coffee event, researchers and students with little or no background in machine learning (ML) are invited to join and ask how ML can be applied in their domain of work. ML+Coffee offers a casual and social atmosphere where ML practitioners can problem-solve with one another. If interested, please contact the ML+X leadership team with a short description of the problem and dataset. For additional context, check out some of the previous projects discussed at ML+Coffee.
|
|
|
|
NMDSI AI Ethics Symposium: From Policy to Practice
November 21, 8:00 a.m. - 6:00 p.m.; Marquette University - Alumni Memorial Union
This event will have two keynote addresses from Dr. Alondra Nelson, former deputy assistant to President Joe Biden and director of the White House Office of Science and Technology Policy, as well as Dr. Casey Fiesler, associate professor of information science at the University of Colorado Boulder. The event will be hybrid. Learn more and register via Eventbrite.
|
|
|
|
ML+X Nexus: Crowdsourced ML and AI Resources
Nexus is the ML+X community’s centralized hub for sharing machine learning (ML) and AI resources, designed to make the practice of ML/AI more connected, efficient, accessible, and reproducible. It is intended to host a wide range of resources (original or external), including educational materials, recorded talks across campus, model use guides, datasets, EDA case studies, and more. While practitioners can use Nexus to expand their expertise, educators and researchers will find it useful for sharing ML knowledge and procedures, reducing redundancy, and improving learning outcomes. Visit ML+X Nexus to begin expanding your knowledge, or visit our How to Contribute page to share a useful ML resource from your field.
|
|
|
|
|
|
Data Science Institute Managing Editor
Apply by November 11 – Sociological Methods and Research (SMR) is a top ranked methodology journal in the social sciences. SMR receives over 200 new submissions and publishes about 40 articles across four print issues per year. The student in this position will serve as managing editor. Visit the job posting from the Data Science Institute to learn more and apply.
Requirements & Qualifications
- Graduate student in any field interested in publishing or Open Science, or who is on a professional or management trajectory
- Knowledge of and/or interest in academic publishing
- Superb written communication skills
- Basic understanding of the research process, incl. quantitative and qualitative methods
- Experience with basic data analysis
Responsibilities
- Manage the workflow of submissions, reviews, and publication
- Communicate in writing with editor, publisher, authors, and reviewers
- Screen manuscripts for adherence to submission requirements
- Ensure adherence to SMR’s Open Science policies and improve journal processes and policies
|
|
|
|
Apply by November 11 – This year-long teaching program is for grad students and postdocs in the biosciences (or connected fields) pursuing careers that include college-level teaching. Fellows get hands-on experience with research-based, inclusive teaching and serve as instructors (not TAs!). Through coursework and practical teaching experience in an undergraduate course, Fellows participate in a supportive community of colleagues and collaboratively develop innovative and effective ways to teach science.
Attend the virtual info session on Thursday, November 7, from 11:30 a.m. - 12:15 p.m., on Zoom, to learn more. Find the Zoom link, the application, and more information on the Scientific Teaching Fellows webpage. Questions? Email Dr. Cara Theisen (chtheisen@wisc.edu). Requirements
- Interest in and enthusiasm for teaching, and a willingness to try new teaching approaches
- Your major field of study involves the life sciences in some way
- Preferred experience in teaching experience, research, and mentoring experience
Responsibilities
- Serve as an instructor and gain firsthand experience designing and teaching a first-year biosciences course
- Develop instructional materials that encourage active learning, and test them in an authentic teaching setting
- Experience all aspects of teaching, from developing learning outcomes to managing a classroom environment to grading
|
|
|
|
Bioinformatics Scientist Position at the Cancer Informatics Shared Resource (CISR), UW-Madison
Apply by November 10 - The incumbent will be part of the Cancer Informatics Shared Resource (CISR) within the Department of Biostatistics and Medical Informatics, which supports research within the University of Wisconsin Carbone Cancer Center (UWCCC). CISR supports scientists and clinicians in the processing and analysis of various next-generation sequencing, proteomic, metabolomic, imaging, and clinical data. Day-to-day operations will include analyzing data, organizing results, giving presentations of results to collaborators and colleagues, assisting with publication preparation, and grant writing for cancer-related research. Visit the job posting from the School of Medicine and Public Health to learn more and apply.
Responsibilities
- Identifies research problems and designs complex research methodologies
- Performs and supervises research, and prepares and presents results for presentations to professional organizations or for scholarly publications
- Attends and assists with the facilitation of scholarly events and presentations
Requirements & Qualifications
- PhD in statistics, computer sciences or related discipline
- Minimum of two years of bioinformatics experience: post graduate work and/or professional experience
- Experience in the data analysis of at least one: Single-cell RNA-Seq, Epigenetics (ChIP-Seq, CUT&RUN, CUT&Tag), Variant detection (SNV, copy number, and structural), Multi-omic integration, Bulk RNA-Seq, Proteomics, Metabolomics, Chromosome organization (4-C / Hi-C), Spatial scRNA-Seq or other -omics
|
|
|
|
New Research Topics: NVIDIA Academic Grant Program Accepting Proposals
Apply by December 31 - NVIDIA’s Academic Grant Program is calling for research proposals to advance work in three new interest areas: Data Science, Graphics and Vision, and Edge AI. NVIDIA will continue accepting submissions for projects related to Simulation and Modeling and Gen AI and LLMs. For more information, please see the NVIDIA FAQs webpage.
New areas of interest:
- Data science submissions can include data processing, operational research and route optimization, and graph neural networks
- Graphics and vision submissions can include augmented and virtual reality, ray tracing, rendering, and AI for graphics
- Edge AI submissions can include robotics, autonomous vehicles, 5G/6G, smart spaces, and federated learning
|
|
|
|
|
|
|
DATA VISUALIZATION OF THE WEEK
|
|
|
|
"Where climate change poses the most and least risk to American homeowners"
Keys and Philip Mulder, of the Wisconsin School of Business, were concerned that millions of homeowners throughout the country could face a slow and undetectable re-pricing of their homes. They tested whether homes vulnerable to flooding were bought and sold differently compared to ones on safer ground. Their methods involved analyzing a million Florida home transactions to see how climate change impacts home values. Unfortunately, they found that the gap between riskier properties and safer ones is growing.
2023 may turn out to be the climate tipping point in the real estate market - can you spot it in the graph? Homes that face a greater flood risk (orange) have a much lower home price appreciation compared to homes with low flood risks (blue). The difference in appreciation amounts to roughly $20,000 on a $400,000 house.
|
|
|
|
Data Science Updates is a collaborative effort of the Data Science Institute and Data Science Hub.
Use our submission form to send us your news, events, opportunities and data visualizations for future issues.
|
|
|
|
|
