+Data Science (+DS) is a Duke-wide program, operating in partnership with departments, schools, and institutes to enable faculty, students, and staff to employ data science at a level tailored to their needs, level of expertise, and interests. For more information, please visit our website at https://plus.datascience.duke.edu.
New Spring vLEs: The Proposal Studios
The proposal studio virtual learning experience (vLE) concept is newly launching in spring 2021, with the goal of assisting Duke investigators with proposal development in health data science, and in sharing experiences with the broader Duke community. The series is co-hosted by Duke AI Health and the Duke+Data Science (+DS) program.
In each one-hour vLE, small teams of Duke investigators will discuss their proposal concepts with data science experts. Attendees will learn more about relevant analysis methods. But more broadly, attendees will also be introduced to approaches for early-stage concept ideation and refinement processes for health data science projects, as well as how to approach proposal development and submission for external funding agencies.
Anyone in the Duke community is welcome to attend, and we especially encourage Duke early-stage investigators, postdocs, and trainees to join us.
Proposal Studio on Structured Data Analyses, Part 1
Monday, March 29 | 11:00 AM – 12:00 PM
Proposal concepts will include traumatic brain injury in Tanzania, claims analysis for Ghana/Nigeria, and hip fracture readmissions, hosted by Matt Engelhard, Ben Goldstein, and Michael Pencina. Register at https://training.oit.duke.edu/enroll/common/show/21/175410
Proposal Studio on -Omics Analyses
Monday, April 5 | 11:00 AM – 12:00 PM
Proposal concepts will include genomic analysis related to sickle cell anemia, lifestyle intervention adherence, and transplant optimization, hosted by David Page, Svati Shah, and Michael Pencina. Register at https://training.oit.duke.edu/enroll/common/show/21/175408
Proposal Studio on Novel Data Sources and Platform Development
Wednesday, April 7 | 1:00-2:00 PM
Proposal concepts will include a repository for global injury, cardiovascular risk in pregnancy, and inpatient hypoglycemia, hosted by Ricardo Henao and Michael Pencina. Register at https://training.oit.duke.edu/enroll/common/show/21/175411
Proposal Studio on Structured Data Analyses, Part 2
Wednesday, April 21 | 12:00 – 1:00 PM
Proposal concepts will include AI-based pulmonary function testing, sickle cell severity, long-term follow up for diabetic retinopathy, and ovarian cancer care recommender, hosted by Samuel Berchuck and Michael Pencina. Register at https://training.oit.duke.edu/enroll/common/show/21/175412
Upcoming Virtual Learning Experiences
Recommendation Systems and the Surveillance Economy
Tuesday, March 23 | 12:00-1:00 PM
Sarah Rispin Sedlak and Akshay Bareja
Recommendation systems—such as the algorithms powering Netflix, suggesting jobs to apply for, and curating Facebook feeds—are powerful tools that can help users navigate an overwhelming array of choices. However, these systems can have negative side effects if left unchecked. In this vLE, we will introduce you to a few popular recommendation-system algorithms, how they work, and discuss how their use may promote homogenization and polarization among target audiences. These effects, while lucrative to service providers, can have negative social consequences. Register at https://training.oit.duke.edu/enroll/common/show/21/175400
Generating Computational Three Dimensional (3D) Geographies
Wednesday, March 24 | 4:30-5:30 PM
Augustus Wendell
This workshop presents an ongoing body of work in computational 3D modeling. The topic revolves around the simulation and rendering of spatial geographic forms, in particular landmasses, landforms and large scale geologic features (mountain ranges). The body of work exists both as computational data sets and large scale visual art works. Affordances in computational methods involving the generation of specific 3D formal language and the processing of ultra high resolution imagery will be demonstrated and discussed. Register at https://training.oit.duke.edu/enroll/common/show/21/175401
Photogrammetry Project Showcase
Friday, March 26 | 3:00-4:00 PM
Hosted by Augustus Wendell
Photogrammetry is a computational method to build 3D models of spaces and objects from collections of photographic images. In this showcase, faculty from Duke University, NC State University and Wake Tech Community College will be presenting overviews of photogrammetry projects. Project topics and scales include: Archeological excavations and artifacts, medieval fortification ruins, prop models for real time gaming, and organic objects for real time VR experiences. Register at https://duke.zoom.us/meeting/register/tJclfuurqDojHtUv9QDm1Hl_KYOymaV8ua0N
An Introduction to Topic Modeling in R
Wednesday, March 31 | 12:30 – 1:30 PM
Chris Bail
This seminar will discuss one of the most popular techniques for automatically identifying latent or hidden themes in text: topic models. We will begin with a brief, high-level introduction to Latent Dirichlet Allocation but spent most of the hour discussing how to write code to perform topic modeling on a corpus of political statements. This webinar will cover both conventional LDA as well as Structural Topic Modeling— a more recent technique that employs meta-data to improve classification of documents according to latent themes or topics. This course assumes a basic working knowledge of R, and the content covered in an earlier “Introduction to Text Analysis” webinar that covers text preprocessing and creating document-term matrices. https://training.oit.duke.edu/enroll/common/show/21/175409
Applying Deep Learning to Biological Sequence Data (A two-part basic sciences session)
Wednesday, April 7 & Thursday, April 8 | 4:30-5:30 PM
Akshay Bareja
Recurrent neural networks (RNNs) are a class of neural networks that can process sequential data, such as text. RNNs have been successfully applied to many natural language processing tasks, including text generation, classification, and translation. In this two-part vLE, we will first introduce you to RNNs and their specific application to biological sequence data. In the second part, we will demonstrate how to build an RNN using PyTorch that can predict protein function based on amino acid sequence data.
Part 1: What is a recurrent neural network? Register at https://training.oit.duke.edu/enroll/common/show/21/175402
Part 2: Implementing an RNN to predict protein function from sequence. Register at https://training.oit.duke.edu/enroll/common/show/21/175403
Coming Soon: Duke Machine Learning Virtual Summer School
Registration will be opening soon for the 2021 Duke Machine Learning Virtual Summer School (MLvSS), which will be held in a virtual format June 14-17, 2021.
The 3.5 day curriculum in the MLvSS is targeted to individuals interested in learning about machine learning, with a focus on recent deep learning methodology. The MLvSS will introduce the mathematics and statistics at the foundation of modern machine learning, and provide context for the methods that have formed the foundations of rapid growth in artificial intelligence (AI). Additionally, the MLvSS will provide hands-on training in the latest machine learning software, using the widely used (and free) PyTorch framework. The MLvSS is particularly well-suited to members of academia and industry, including students and trainees, who seek a thorough introduction to the methods of machine learning, including interpretation and commentary by respected leaders in the field.
The 2021 MLvSS will be led by a trio of machine learning experts at Duke University: Professors Ricardo Henao, David Carlson, and Timothy Dunn. They will be joined with lectures by other Duke professors and by the founding director of the machine learning schools, Lawrence Carin. Hands-on training with software will be provided by Duke graduate students who have extensive experience with these tools, and teaching assistants from the Duke AI Health Fellowship program will be available for assistance throughout the course.
Eight Duke Machine Learning Schools have been presented since 2017, reaching hundreds of participants from academia and industry and including international audiences at the SingHealth/Duke NUS Medical School and the Duke Kunshan University campus.
