Thank you for your interest in undergraduate research with us!
Please fill out the Application Form when ready to join the team!
This page should help give you the key information needed to explore doing research with us. If after reading this you have any questions please reach out to us via email (skunkworks@mse.wisc.edu).
The Quick Overview:
- Join the Information Sessions December 4th and 5th!
- Anyone who wants to join the Skunkworks is welcome and the main requiredment is the time commitment (~10 hrs/week) to participate.
- If you are interested in the on-boarding ML4ER course there are no background requirements and if you fill out the application form you can safely assume you are enrolled and will get followup information for the Spring ’24 course by Dec. 19th.
- For research students we will do our best to find a project match but cannot guarantee project availability. We will reach out to applicants by Dec. 19th to confirm applications and schedule an informational interview.
- For application submissions after Dec. 17th we will followup in ~1 week to let you know about any continuing openings and will add you to our email list to stay up to date for future opportunities.
Who we are.
We call our organization the Informatics Skunkworks. The Skunkworks is a group that leads mentored group research projects for undergraduate students. Research mentors are based primarily at UW-Madison but students are encouraged to participate from any location, background, and level of experience. In addition to the group based research projects led by the Skunkworks, we also hosts community wide events and resources to support student researchers in their specific research projects. Our mentors are typically postdocs, staff, and faculty with extensive experience. The online course and project will be led by Benjamin Afflerbach, with over 7 years of experience in machine learning for materials.
What it means to participate on a research project.
Participation involves attending weekly research group meetings, communicating results through slide decks, working independently and in small groups on specific research tasks (clean a dataset, train/assess an ML model, compare multiple model types). In addition, participants will construct a “Final Slide Deck” at the end of each semester to summarize their progress. Regular weekly activities occur from September to December in the Fall and January to May in the Spring. Overall activities are expected to have a time commitment of around 10 hours/week. We encourage students to join us for multiple semesters but this is not required.
What to do if you aren’t quite ready to participate in research.
We have many students reach out who are interested in machine learning, but are missing the background knowledge and skills to immediately jump into a research project. To support those students we lead an on-boarding “Introduction to Machine Learning for Engineering Research” (ML4ER) course each Fall and Spring semester. This course is separate from the research projects but has the goal of preparing students to participate in future semesters. For example you could take the course in the Fall and then join a research project in the Spring.
How to join a project or the “Introduction to Machine Learning for Engineering Research” ML4ER course.
Both projects and the ML4ER course run from September to December in the Fall, and January to May in the Spring. Recruitment for both is focused around the start of each term (Sept. and Jan.) where project groups and course enrollment are established. To apply to join students can submit an application via the Steps to Join section. Additionally, we will host several information sessions at the start of each semester to help give out information and answer questions. The details for these will be posted on the website as well as sent out through all of the social media channels below.
Join a Skunkworks Research Project
For UW-Madison students there are a number of current projects you can join. For students outside of UW-Madison we are also offering a specific generative machine learning project that we invite you to join! We want to get students involved in authentic mentored research experiences, see the tabs below for all the details to get started!
Projects are setup to generally take 1-2 years to complete and we encourage students to participate across multiple semesters to see the full scope of research. For a few examples of past projects see our project highlights.
- Steps to Join
- Background Requirements
- Participation Expectations
- Academic Credit (UW-Madison)
- Payment
- Review the Information and Requirements listed on this page.
- (UW-Madison Students) Check the Project Page for a list of active projects. Note that most projects recruit around the start of the Fall and Spring semesters.
- (non UW-Madison Students) We are offering a remote project on Generative Machine Learning and have setup a separate page with project details.
- Fill out the Application Form to let us know about your background, interests, and any specific projects you are interested in. We will help facilitate an interview with the project mentor.
- Attend an information session to ask questions (or send any specific questions to Ben Afflerbach (bafflerbach@wisc.edu). Details for each semesters specific dates and times will be posted on the website and sent out through our social media channels!
- Project mentors will followup to confirm participation and schedule group meetings.
- A computer and internet connection
- Ability to explain and execute common machine learning models tasks such as:
- Data Cleaning
- Featurization
- Models Assessment
- Hyperparameter Optimization
- Experience with programming languages such as Python and R (individual projects may use others as well).
- Approximately 10h/wk work on your skunkworks projects.
- Participation in project meetings, including regular reporting (typically a slide deck each week).
- A final report slide deck summarizes your work for the semester.
- Any additional requirements associated with obtaining academic credits, if you wish to do that. Skunkworks participation is typically 2 credits as it is about 10h/wk for about 10 weeks (~100h). The first few weeks and last week are typically not active due to delays starting and avoiding conflict with final projects/exams.
We encourage but do not require:
- Participation in Skunkworks community events, e.g., talks, workshops, and the end of semester “All-Hands” meetings.
- Engaging with the skunkworks community through Slack, etc.
Credits can be obtained through a skunkworks project by signing up for an independent study with whoever is advising your project. As many students work with Dane Morgan, here are his more detailed guidelines, but these may differ for different advisors.
Dane Morgan’s credit and grading approach:
Credits are typically obtained through signing up for MSE 401.
Grades will be based on regular slide decks from each meeting and a summary slide deck I will ask for at the end of the semester. The basic grading rubric is below, although I reserve the right to alter this for a given individual to address particular situations:
- Attended all meetings or had an excused absence, Submitted all slide decks for all meetings and final slide deck, and showed ~4x(# credits) hours of work/wk on average in those slide decks = A
- Missed some meetings and/or slide decks and showed somewhat less than ~3x(# credits) hours of work/wk = B
- Missed many meetings and/or slide decks and showed significantly less than ~3x(# credits) hours of work/wk = C
- Missed most meetings and/or slide decks and showed far less than ~3x(# credits) hours of work/wk = F
Here are some useful items related to the slides deck expectations:
- Make sure to create a slide deck each meeting (even if you miss the meeting) that includes what you did, any problems, and plans. Make the slide deck with simple bulleted text so it is understandable without you being present.
- Please include in each slide deck.
- First slide: A title slide with your name, date, the project, and course name and # and number of credits.
- Second slide: A brief summary of hours you worked as a list of date (mm/dd/yy): times (e.g., 1pm-3pm). Make this list cumulative for the semester (i.e., just add the new times each week and keep all the old times).
- Third slide: A bullet summary of what you view as the main accomplishments of the reporting period (all semester for final slide deck). Note these can include learning things, fixing bugs, etc. They should approximately represent where your time went.
- Please include in each slide deck.
- Make sure to put in ~4h/wk per credit or you can quickly get so behind you cannot make up the time.
- Consider meeting with your groups at least 1-2 times per week to create regular checkins.
- Name slidedecks as follows: LASTNAME_YYYY-MO-DD_PROJECT_KEYWORDS, e.g. Morgan_2019-11-27_Skunkworks_Notes on Slide Formatting
Final slide deck: For everyone taking Skunkworks for some kind of credit please complete a final report slide deck summarizing the major achievements for the semester. This is typically 10-20 slides, depending on your scale of effort (e.g., 1-3 credits). Feel free to borrow from previous slide decks. Please follow the formatting for regular slide decks but give a summary of hours by week on the 3rd slide.
- Please put on gdocs in usual place for your project slides and email to me as an attachment. Due by 12 midnight last day of classes.
- Name should be LASTNAME_FINAL_YYYY-MO-DD_PROJECT_KEYWORDS (KEYWORDS are optional).
Credit requirements: https://kb.wisc.edu/vesta/page.php?id=24558
Generally, UW-Madison will follow the federal credit hour definition: one hour (i.e. 50 minutes) of classroom or direct faculty/qualified instructor instruction and a minimum of two hours of out of class student work each week for approximately fifteen weeks, or the equivalent engagement over a different time-period.
Alternatively, a credit hour will be defined as the learning that takes place in at least 45 hours of learning activities, which include time in lectures or class meetings, in-person or online, laboratories, examinations, presentations, tutorials, preparation, reading, studying, hands-on experiences, and other learning activities; or a demonstration by the student of learning equivalent to that established as the expected product of such a period of study.
In all cases, learning in for-credit courses is guided by a qualified instructor and includes regular and substantive student-instructor interaction.
Students can be paid for their participation if there is financial support from the project advisor available. This is rare and depends on the project, but please ask when signing up if this is something you would like to pursue. Please also see fellowships information in the Resources page for more information on ways to raise funds.
Join The Skunkworks On-boarding Course (ML4ER)
The Introduction to Machine Learning for Engineering Research (ML4ER) course is a one semester on-boarding course which we teach each semester to help new students get up to speed and hopefully join a research project the following semester. It is open to anyone from any location and you just need an internet connection and computer to participate!
- Steps to Join ML4ER
- Background Requirements ML4ER
- Participation Expectations ML4ER
- Academic Credit (UW-Madison) ML4ER
- Review the Information and Requirements listed on this page as well as the course overview page.
- Fill out the Application Form.
- Note: If also wanting to enroll for academic credit for UW-Madison students you will also need to enroll in MSE 401 which is the course listing associated with the Skunkworks. See the credit tab for details.
- Attend an information session to ask questions (or send any specific questions to Ben Afflerbach (bafflerbach@wisc.edu). Details for each semesters specific dates and times will be posted on the website and sent out through our social media channels!
- We will followup with confirmation and the first course meeting will be the week of Sept. 18th
- A computer and internet connection
- Some familiarity with programming will make a smoother introduction, but not required. Specifically we will be using Python throughout the course for all of our hands on activities.
- Approximately 6-9 hours a week time commitment.
- Attendance at one course meeting a week
- Submission of weekly activity homework assignments and an end of semester final project.
We encourage but do not require:
- Participation in Skunkworks community events, e.g., talks, workshops, and the end of semester “All-Hands” meetings.
- Engaging with the Skunkworks community through Slack, etc.
Credits are typically obtained through signing up for MSE 401 which is available for UW-Madison students. Include in your application that you are interested in enrolling for credits and we will follow up with you about details for enrolling.
Please see the course overview page for a syllabus which includes a breakdown of the grading.
For Mentors (grad student, Faculty, postdoc):
For those with more experience and/or seniority it may not be appropriate to be a traditional Skunkworks participant, who is typically seeking research experience on a project set by someone else. However, we would love to have you involved. Here are few ways you could be involved, somewhat in order of time commitment.
- If you are just interested in hearing about upcoming community events and announcements, join the Google Group email list and our LinkedIn community page which is used in the same way.
- If you are interested in contributing more towards the community, please join our slack channel to both learn about ongoing events and potentially answer questions and informally support others research: https://uwcmg-informatics.slack.com/signup
- Give a talk on a relevant topic at some point to team members. Let Dane Morgan know if you want to do that and he will get it scheduled.
- Lead a research team on a project. This would require having a project ready that is undergraduate appropriate and accessible. If you think you might want to do that please discuss with Dane Morgan. It would likely take ~3h/wk.