In June 2023, Stanford Research Computing (SRC) selected Annie K. Lamar to be the inaugural Stanford Research Computing Humanities Research Fellow. The first in her family to attend college, Lamar earned her Ph.D. in Classics and was a Data Science Scholar in the Stanford Data Science program. She also holds bachelor’s degrees in Classics and Computer Science from the University of Puget Sound, and a master’s degree in Education Data Science from Stanford.
The annual Summer Quarter fellowship is a way for Research Computing to reach out and make our resources available to graduate students in the humanities or social sciences who are doing original digital research. Applicants of all experience and degree levels are welcome to submit a proposal for a research project at any stage of ideation and planning. We are especially interested in funding projects from early career scholars, master’s students, and individuals from underrepresented backgrounds.
Fellows receive a $5,000 stipend, office space in Polya Hall, and assistance from SRC staff to implement their research in an HPC (high performance computing) environment. At the end of the summer, fellows give a formal presentation about their project.
Annie K. Lamar was selected to be the inaugural fellow based on her proposal to apply computational methods to the study of ancient literature.
“We were excited to have Annie because her project was strong in both the humanities and computation,” said research data facilitator Brad Rittenhouse, Ph.D., who initiated and administers the fellowship as part of his research data facilitator role in Stanford Research Computing. “Annie has contributed new techniques and a new way of asking questions about some of the oldest and most studied writings in the canon of ancient literature. And our hopes were confirmed: her work here last summer showed the benefit in doing this type of humanities research in an HPC environment.”
Recently, Rittenhouse and Lamar connected for a discussion about her Summer 2023 fellowship experience.
Rittenhouse: Annie, in a nutshell, how would you describe your project?
Lamar: My dissertation research is based on ‘data-intensive humanities,’ which uses new methods of neural network analysis and data-driven methods for analyzing poetry. In particular, I’ve been studying a new computational approach to the Homeric formula: repeated phrases that shape the meaning, meter, and performance of Homer’s works such as the Iliad and the Odyssey.
Using results from a neural prediction model, a sophisticated data mining application that imitates the function of the brain to detect patterns in data sets, I aimed to identify phrases that do not survive in Homer’s works, but for which the model finds similar meaning and place.
This project aligns with my goals as a ‘small data’ data scientist. In all my research pursuits, I work with datasets that are less than five percent the size of the huge datasets used to train most world-class artificial intelligence (AI) models. Creating models that perform well on small and specialized datasets will continue to gain importance in all fields as we begin to ask AI to answer more specific, complex, and nuanced questions. I hope to use my experience as a small-data AI researcher and a classicist to promote understanding of AI and data ethics in the humanities.
Rittenhouse: What tools, computing platforms, programming languages, etc., did you use to develop this?
Lamar: Most of my codebase is written in Python, which has many libraries available for natural language processing (NLP) tasks. For example, PyTorch and TensorFlow are two of the most popular libraries for machine learning research. My work in the Sherlock cluster, the high performance computing environment I used at Stanford, was done entirely in the command line.
Rittenhouse: What support/help/advice did you receive from our team members? Are any of these things you would not have had without working with Stanford Research Computing?
Lamar: The most difficult part of most Computer Science tasks is not necessarily learning how to do things correctly, but rather what to do when something goes wrong. The SRC team has really fantastic documentation for independent learners on their website–but you’re still guaranteed to make mistakes or encounter new errors! A valuable part of this fellowship was having a mentor who taught me the troubleshooting process for HPC, essentially teaching me how to learn effectively in this field.
The entire SRC team was incredibly supportive. Every day, people in the office would ask about my project and offer their help. Having people available to troubleshoot with me, in person, everyday was an irreplaceable experience.
Rittenhouse: What did an average day look like while working on your project for the fellowship? In addition to giving a “slice of life”, can you also speak to your ability to fit the fellowship into your existing research portfolio, research plans and trips you may have had for the summer, etc.?
Lamar: Since the fellowship was supporting my dissertation research, it not only gave me the opportunity to build new skills but also to make substantial progress on my dissertation! The skills I built during the project were also widely applicable to my other research projects and teaching. For example, I was able to show my interns how to use FarmShare and make more informed recommendations for other research teams about computational resources. It was also straightforward to fit the fellowship into my existing plans; I started my work on the fellowship once I returned from teaching in Greece in June.
A typical day would include checking in with Brad Rittenhouse, the founder and lead mentor for the fellowship program, about the status of my project and what I wanted to accomplish that day. He would teach me a set of new skills related to HPC, send a link to a tutorial, or set up a meeting with another SRC team member. I would work on my task while asking questions. At the end of the day, we would reflect on what I accomplished and how it fits into my overall goals for the summer. Having access to Polya hall and a designated desk nearby the members of the SRC team was a real benefit of the program.
Rittenhouse: What have been the benefits/effects of this opportunity on your academic research and/or your professionalization outside of academia?
Lamar: There’s so many opportunities on campus for graduate students to mentor others, but this is an opportunity for graduate students to receive individualized mentorship themselves for a unique skill set rarely taught in Humanities environments. The skills I built this summer–understanding HPC systems, writing SLURM scripts, working in the command line–have made all my research easier. In addition, being familiar with Stanford’s resources has reduced (or eliminated) my need to pay for outside computing resources.
One outcome I did not expect is how much this fellowship would facilitate me connecting with other researchers outside my discipline, including people in Linguistics, Neuroscience, and Music. Although our fields and projects are incredibly different, we all have a shared common core of HPC skills. Because of that, we can offer each other method-specific advice or review that is difficult to receive from people outside the HPC world. By necessity, this fellowship also improved my skills in code collaboration environments (e.g. GitHub). Being an adept digital collaborator is a crucial skill for modern computational Humanists, and it’s–for obvious reasons–a hard skill to learn on your own.
Finally, being a student on this fellowship greatly improved my ability to teach these skills. Because of this fellowship, I’ll be able to teach my own classes on HPC. Brad Rittenhouse started this fellowship program in part because his own experiences learning HPC were so difficult—he wanted graduate students at Stanford to have the opportunity to learn in a supportive and structured (and funded!) environment. I’m so grateful to have had this opportunity. Next fall I will start as Professor at UC Santa Barbara. Maybe we can even start a similar fellowship for graduate students there!
More Information
Annie K. Lamar: Resumé | Projects | Publications
