Prospective PhD Students

I take PhD students through two programs at UCSD. Please apply to both:

Both applications go through the UCSD graduate portal. The GRE is not required. Fee waivers are available (US applicants; international applicants to CSE), and requests are due before the application deadlines.

After you apply, please fill out the Prospective PhD Students form and send me a short email at berk@ucsd.edu to let me know.

Thank you for your interest in my group! If you’re interested in working together, please skim the list below or my research page to see what I’m working on these days. Find some papers that look interesting and read their abstracts. Pick one paper and read it fully. If you’re still interested, complete one of the forms below:

• Prospective Postdoctoral Researchers
• Prospective PhD Students
• Visiting Student Researchers

Here are some papers that I am most excited about:

Interaction

• Classification with Conceptual Safeguards
• Participatory Personalization in Classification
• Actionable Recourse in Linear Classification

Alignment

• Selective Preference Aggregation

Interpretability

• Learning Optimized Risk Scores
• Concept Bottleneck Large Language Models

Reliability

• Predictive Multiplicity in Classification
• Regretful Decisions under Label Noise
• On the Epistemic Limits of Personalized Prediction

Safety

• When Personalization Harms Performance
• Responsiveness Verification: Will Predictions Change? How Much? How Often?
• Operationalizing the Search for Less Discriminatory Alternatives in Fair Lending