Data Science @ UW
Connecting people with data science research, education, and collaboration at UW-Madison

ML+X Celebrates Two Years of Community Building Around Machine Learning

Posted on

by Chris Endemann (with some assistance from ChatGPT)

Two years ago, the ML+X community was launched with a clear mission: to bring together students, researchers, and industry professionals who share an interest in using machine learning (ML) methods to advance their work (+X).  Given the ever-evolving landscape of ML and its widespread use in nearly every scientific field and industry, this goal seemed (and still is!) critically relevant. By forming bridges across domains and providing opportunities for people of different backgrounds and experience levels to engage with ML, our aim has been to make the practice of ML in Madison more connected, accessible, efficient, and reproducible. As ML+X celebrates its second year anniversary, we reflect on our journey of fostering communal learning and collaboration in the realm of ML. Looking ahead, we discuss our future plans to further advance our impact.

Community Growth & Demographics

Over the past two years, the ML+X community has grown rapidly, welcoming 331 new members in 2023 (677 in total!). This steady influx is a testament to the community’s relevance in the landscape of ML. Our community represents a diverse mix of professionals and enthusiasts from various fields, including Biomedical Sciences, Chemistry, Agriculture, Atmospheric and Oceanic Sciences, Psychology, Data Science, and more. Members include graduate students, academic staff, researchers, undergraduate students, and industry professionals, creating a diverse spectrum of expertise. This diversity extends into a learning environment that accommodates individuals at different stages of their ML journey. From seasoned experts with over a decade of experience to those taking their first steps, ML+X cultivates a culture of collaborative learning. Valuable insights are shared through meaningful conversations and shared experiences, enriching the learning experience for all members. As we celebrate this growth, we recognize its impact on the ML+X community’s overarching mission. This vibrant community, with its diverse membership, sets the stage for continued collaboration and knowledge exchange.

ML+X Events

In-Depth Forums and Interactive Discussions

The sixteen ML+X Forums held monthly over the past two years have served as focal points for in-depth discussions covering a spectrum of ML topics and applications. These discussions have ranged from exploring multimodal learning in functional genomics to delving into protein engineering using large language models, and addressing bias and fairness in ML. For a comprehensive list of our past forums, please refer to the ML+X Forum archive. Starting this fall, we have introduced the recording of select forums (with presenter consent) to enable members to catch up on any missed sessions (see ML+X Forum playlist).

During the monthly forums, our community has embraced a shift toward interactive and open discussions. This two-way communication emphasizes engagement and collaboration, fostering an exchange of insights between presenters and the audience. The collaborative atmosphere is integral to upholding the standard of communal learning within ML+X. Presenters are not expected to have all the answers; instead, the emphasis lies on sharing methods and actively seeking additional insights and ideas from the audience.

To provide a glimpse into the enriching experience of presenting at ML+X Forums, here are thoughts from two of our past presenters:

Rohan Sonthalia, an undergraduate student using deep learning to extract insights from DNA sequences, shares, “My experience as a presenter at the ML+X Forum allowed me to interact with like-minded individuals. Engaging in conversations during the presentation was not only interesting but also proved helpful in overcoming roadblocks that I faced in my work.” 

Mike Killoran, a Senior Scientific Researcher at Promega with a background in biochemistry and over 10 years of experience in ML, adds, “I was impressed with the curiosity and enthusiasm of the group during the presentation and was very grateful to connect with scientists at UW who think about solving similar challenges in protein design and engineering.”

Whether you’re a seasoned professional or just starting your journey in ML, the ML+X Forum provides an inclusive platform to discuss your projects or proposals. All experience levels are welcome to present and seek valuable feedback and ideas related to their ML methods or project challenges. If you’re currently working on an ML project or about to embark on one, we encourage you to fill out the presenter sign-up form and join our collaborative community.

Monthly Social – ML+Coffee

The morning after each monthly ML+X Forum, the community hosts a social event known as ML+Coffee (coffee/tea provided, laptops encouraged!). ML+Coffee provides a casual format for community members to form connections from a diverse member base, discuss the latest news in ML, troubleshoot and diagnose ML problems, engage one-on-one with ML+X Forum presenters, and share learning resources. While this event is intentionally more casual, the conversations at this event consistently yield insightful perspectives and ideas that members can integrate into their work. To offer insight into the collaborative nature of ML+Coffee, we asked a couple of our regular attendees to comment on their experiences:

Rafeh Qureshi, an economics Ph.D. student shares, “As someone whose research involves ML tools, but whose background had been primarily from other fields, ML+Coffee provided a unique and much needed community of experts whose experience I was able to draw on for questions I had and ideas I had been wondering about. Whether through asking direct questions or just listening to discussions about research and ideas in various fields using ML, the forum has been very helpful in helping me expand my understanding of ML and introducing me to a wider set of computational tools. “

Mark Banghart, a data scientist in the department of Ophthalmology shares, “When I started using ML, the ML+Coffee gatherings allowed me to ask questions and learn from other practitioners. The suggestions on approaches and resources assisted in my development. The ML+Coffee is a friendly and accepting environment to exchange experiences and ask questions about ML.”

2023 Machine Learning Marathon

The inauguration of the 2023 Machine Learning Marathon marks a new milestone in our community’s journey. This extended hackathon-style event is a contest where teams of practitioners compete to secure the highest leaderboard ranking across different Kaggle ML challenges. Beyond the intrinsic rewards of skill enhancement and community building, the stakes are heightened by the prospect of a cash prize for the winning team, generously provided by our sponsors. The beauty of this marathon lies not only in its diversity but also in its inclusivity—projects cater to varying skill levels, ensuring that every participant finds a meaningful challenge to tackle. In addition, one distinctive feature of the Machine Learning Marathon is the allocation of experienced advisors to each team, providing invaluable guidance throughout their project endeavors. Looking ahead, ML+X aspires to further elevate the marathon experience by incorporating local challenges and datasets, with the ultimate goal of featuring a UW-Madison project on Kaggle. We enthusiastically invite readers to suggest potential projects by emailing endemann@wisc.edu.

Connections and Collaborations

Thank You, Sponsors!

An exciting and pivotal development for ML+X has been the acquisition of its first sponsors: Epic, Promega, and WARF. Their generosity has been instrumental in our growth, enhancing the quality and reach of our community initiatives. Beyond addressing logistical needs, their financial support has contributed to creating a more inclusive and engaging environment, broadening our audience. Whether it’s covering catering or providing hackathon prizes, these seemingly small details have significantly improved our events, making them more conducive to collaboration. ML+X is immensely grateful to Epic, Promega, and WARF for their unwavering support. Their commitment has empowered us to pursue our goals, facilitating a community where ML practitioners explore challenges, share knowledge, and support each other’s work.

Ethics, Values, Information and Law (EVIL)

ML+X’s collaboration with the Ethics, Values, Information, and Law (EVIL) reading group signifies our commitment to delving into ethical considerations inherent in data and information technologies. The EVIL Reading Group meets online every three weeks (roughly) on Fridays. Past discussions have been both enlightening and timely, covering a range of topics including:

These discussions not only highlight ML+X’s commitment to ethical exploration but also showcase the impactful conversations fostered within our community. By actively engaging with topics that touch upon legal and ethical considerations, ML+X aims to contribute meaningfully to the broader discourse on the responsible and ethical use of ML technologies.

ML Training Opportunities Via The Carpentries

In collaboration with the UW-Madison Carpentries Community, ML+X has helped develop and host several ML workshops throughout the calendar year, including:

Moreover, ML+X has two workshops currently in the incubator phase, namely: (1) Reproducible and Responsible ML and (2) Intro to Deep Learning with PyTorch. We encourage our readers to suggest additional workshops at any time by emailing facilitator@datascience.wisc.edu.

The collaborative efforts of the local Carpentries community, comprising workshop instructors, helpers, and lesson developers from various campus roles, including research support staff, faculty members, and students, make these workshops possible. Their dedication empowers others with the essential tools and knowledge for ML, and we extend our sincere gratitude for their invaluable contributions. For those interested in exploring volunteer opportunities within UW-Madison’s Carpentries community, we invite you to check out the volunteer FAQ page for more information.

Looking for registration details? Upcoming data science and ML workshops are regularly announced every two weeks in the Data Science @ UW Newsletter — be sure to subscribe to receive the newsletter directly in your inbox. You can also check the Training Resources page for workshops currently open for registration.

MadPy

ML+X has forged a budding collaboration with MadPy, a local Madison community dedicated to all facets of the Python programming language. In the past two years, our relationship has seen the exchange of event promotions and fostered cross-pollination between our organizations. Notably, ML+X contributed to a MadPy meetup by hosting a demo on Streamlit, enriching the collective knowledge of both communities. As our collaboration evolves, we are excited to announce our joint venture with MadPy in organizing next year’s Machine Learning Marathon event. This partnership signifies our commitment to fostering a robust and interconnected ecosystem, where Python enthusiasts and machine learning practitioners converge to explore the endless possibilities within these fields. Stay tuned for more collaborative initiatives between ML+X and MadPy on the horizon.

Future Initiatives

Exploratory Data Analysis Guide

As ML+X looks to the future, we are strategically advancing with the development of an Exploratory Data Analysis (EDA) guide. Recognizing that EDA is a cornerstone in any ML pipeline, we emphasize the pivotal role it plays in setting the stage for successful projects. Without a deep understanding of the data, ML projects can become an iterative loop, hindering progress and efficacy. This comprehensive guide is being curated from case studies generously contributed by our community members. It aims to illuminate the nuanced questions and considerations essential for addressing diverse problems across various domains. We’re currently in the process of unveiling a series of “gold standard” case studies, carefully selected to kickstart this initiative. These cases will not only serve as instructive examples but also as foundational references for mastering the art of exploratory data analysis. Stay tuned for more updates soon!

Have an EDA case study to share? If you would like to contribute an EDA example to this guide, please email me at endemann@wisc.edu.

Leadership Expansion and Diversity

As ML+X enters its third year, we are excited to announce our pursuit of an expanded and more diverse leadership team. We invite individuals with diverse expertise and perspectives, especially those from the humanities, social sciences, and environmental sciences, to join us in steering the community towards greater growth and impact. Recognizing that leadership diversity enhances the richness of our collaborative efforts, we believe in the power of varied perspectives to shape the future of ML+X. If you are passionate about contributing to our vibrant community and helping guide its trajectory, we welcome you to join the team. If interested, please email me at endemann@wisc.edu.

Member Hosted Events

One of the hallmarks of ML+X is the community’s participatory nature. We strongly encourage our members to take an active role in shaping the ML+X experience by hosting events. These events can encompass a wide range of formats, from focused workshops to informal discussions and networking sessions. Hosting an event provides you with an opportunity to share your knowledge, interests, or projects with the community while fostering connections and collaborations. If you have a topic you’re passionate about or an ML-related resource you’d like to showcase, please email me at endemann@wisc.edu . Your contributions play a crucial role in diversifying the community’s offerings and creating a more inclusive and engaging space for all members.

Conclusion

As ML+X reaches its two-year milestone, we reflect on a journey marked by diversity, collaboration, and innovation in ML. Our community serves as a hub for students, researchers, and professionals united by a passion for advancing their work through ML methods. From engaging discussions in our forums to the camaraderie of ML+Coffee gatherings, our experiences embody the essence of collaborative learning. Looking ahead, ML+X is committed to making meaningful contributions to the field of ML. Join us in this journey by becoming a part of our community through the ML+X Google Group and Slack channel (#ml-community).

ChatGPT Disclaimer

I experimented with using the free version of ChatGPT to expedite writing this blog. With some carefully crafted prompts and prior information fed into the bot, I found it did speed up the process a little. However, it was far from perfect. ChatGPT tends to overuse buzz words (e.g., for some reason it thinks everything is dynamic) and sounds more like a salesman than an educator. I suppose this makes sense given that it was trained on large swaths of internet text… which is mostly advertisements. In any case, I edited the blog to convey my own voice and ensure certain details weren’t obfuscated. I may use it again to assist my writing, but only when used in conjunction with (1) a careful review process and (2) transparent acknowledgment.