Personal Research Statement
for Ph.D. Programs in Machine Learning

A few people have asked to read my personal research statement for Ph.D. programs in machine learning, so I’ve released my statement here (link). The statement follows the below outline, which I think is a generally good outline for such statements:

1. Describing my overall research interests at the time (first paragraph)
2. My prior research, in order from most to least important/impressive/relevant (several paragraphs)
3. A paragraph on why I was excited to apply to a particular school, including specific faculty names I was interested in working with.

Reading my statement now, I find it fairly cringe-y, and my writing (and humility) has improved a lot since. I wouldn’t imitate my statement super closely, but I do think the high-level structure is good. I ended up getting accepted to around half of the programs that I applied to.

My research interests have also updated significantly, and I now believe that work on machine learning safety is much more pressing than work on improving machine learning capabilities like language understanding. As a result, my research agenda now focuses on safety, and I’d encourage others applying to grad school to consider safety-related topics like interpretability, adversarial robustness, red teaming, backdoor attacks, and aligning models with human preferences.

Best of luck applying to Ph.D. programs!