Ioana Marinescu

I am a PhD student at the NYU Center for Data Science. My broad research interests include meta-learning, reasoning, and priors:

• meta-learning: understanding in-context vs. in-weights learning for pretraining and finetuning
• reasoning: disentangling abstract reasoning from task-specific content
• priors: learning useful priors.

Prior to NYU, I graduated summa cum laude from Princeton with a Bachelor’s degree in Computer Science. I previously worked on inductive bias distillation into neural networks, symbolic regression, and noun compound understanding.