Aaron Mueller is an Assistant Professor of Computer Science at Boston University. His lab develops methods for improving our understanding of the internal workings of natural language systems, such as language models. His research goals are to evaluate and improve the robustness of NLP systems, to develop techniques that allow us to precisely fix model errors in a way that humans can understand without harming general capabilities, and to leverage insights from these studies to improve language model training or fine-tuning methods.

The lab’s research centers on three main directions: (1) causal and mechanistic interpretability methods and frameworks; (2) evaluations of language models inspired by linguistic principles and findings in cognitive science; and (3) developing more effective NLP systems—for example, by proposing improvements to language model training that make them more sample-efficient and performant.

He obtained his Ph.D. in Computer Science at Johns Hopkins University advised by Tal Linzen and Mark Dredze. Before BU, he was also a postdoctoral researcher working with David Bau and Yonatan Belinkov.

Aaron is interested in mentoring a research project that includes at least one of the following:

• Developing strong evaluations of LMs that directly assess generalization and robustness

• Jointly developing methods and applications of model internals–based interpretability

• Sample- and compute-efficient language model training or adaptation methods

• Developing theories of interpretability grounded in causality or differential geometry

If you already have an idea you’d like to pursue, great! If you don’t yet know what you’d like to work on or want to join an existing project, also great!

Candidates should have one or more of the following:

• Previous research experience in machine learning or NLP, as evidenced by a first-authored paper in an ML or NLP venue

• Familiarity with machine learning toolkits and methods, including PyTorch and HuggingFace

• Experience in training, post-training or fine-tuning large-scale neural networks (ideally language models)

The following will also be expected:

• A collaborative approach to scientific research

• A desire to both be mentored and to be a mentor to others

• A desire to share your work (including ideas, results, code, and scientific process) with others in the academic community