Domain Adaptation and Related Areas

Program

Modern Paradigms in Generalization

Location

Calvin Lab Auditorium

Date

Tuesday, Nov. 12 – Friday, Nov. 15, 2024

Back to calendar

8:45 – 9:15 a.m.

Coffee and Check-In

9:15 – 9:20 a.m.

Opening Remarks

9:20 – 10 a.m.

A Discrepancy-Based Theory of Adaptation

Mehryar Mohri (Google Research & NYU)

10 – 10:40 a.m.

Label Noise: Ignorance is Bliss

Clayton Scott (University of Michigan)

10:40 – 11 a.m.

Break

11 – 11:40 a.m.

Does Distribution Shift Matter In the Era of Pre-trained Language Models?

Robin Jia (University of Southern California)

11:40 a.m. – 12:20 p.m.

The Data Addition Dilemma

Irene Y Chen (UC Berkeley)

12:20 – 2:30 p.m.

Lunch (on your own)

2:30 – 3:10 p.m.

How Transformers Learn Causal Structure with Gradient Descent

Jason Lee (Princeton University)

3:10 – 3:50 p.m.

Pseudo-Labeling for Covariate Shift Adaptation

Kaizheng Wang (Columbia University)

3:50 – 5 p.m.

Poster Session and Reception

8:45 – 9:20 a.m.

Coffee and Check-In

9:20 – 10 a.m.

TBD

Nika Haghtalab (UC Berkeley)

10 – 10:40 a.m.

Transfer learning for weak-to-strong generalization

Yuekai Sun (University of Michigan)

10:40 – 11 a.m.

Break

11 – 11:40 a.m.

Distribution shift in ecological data: generalization vs. specialization,

Sara Beery (MIT)

11:40 a.m. – 12:20 p.m.

Transfer learning via local convergence rates of the nonparametric least squares estimator

Johannes Schmidt-Hieber (University of Twente)

12:20 – 2:30 p.m.

Lunch (on your own) and Lunchtime Activity

2:30 – 3:10 p.m.

The Modulus of Transfer

Steve Hanneke (Purdue University)

3:10 – 3:50 p.m.

User-level and federated local differential privacy

Yi Yu (University of Warwick)

4 – 5:30 p.m.

Optional Sunset Hike

8:45 – 9:20 a.m.

Coffee and Check-In

9:20 – 10 a.m.

On Spurious Associations and LLM Alignment

Victor Veitch (University of Chicago)

10 – 10:40 a.m.

TBD

Zachary Lipton (Carnegie Mellon University)

10:40 – 11 a.m.

Break

11 – 11:40 a.m.

Causally motivated robustness to shortcut learning

Maggie Makar (University of Michigan)

11:40 a.m. – 2:30 p.m.

Lunch (on your own) and Table Discussions (click for sign up sheet)

Sign up for a group discussion here (click here)

2:30 – 3:10 p.m.

TBD

Emma Brunskill (Stanford University)

3:10 – 3:50 p.m.

Toward Understanding the Extrapolation of Nonlinear Models to Unseen Domains or the Whole Domain

Tengyu Ma (Stanford University)

3:50 – 4:10 p.m.

Break

4:10 – 5 p.m.

Language-guided Adaptation

Trevor Darrrell (UC Berkeley)

8:45 – 9:20 a.m.

Coffee and Check-In

9:20 – 10 a.m.

Distributional Shifts and Spectral Risks

Zaid Harchaoui (University of Washington)

10 – 10:40 a.m.

Adaptive Data Collection via Autoregressive Generation

Hongseok Namkoong (Columbia University)

10:40 – 11 a.m.

Break

11 – 11:40 a.m.

Scalably Understanding AI with AI

Jacob Steinhardt (UC Berkeley)

11:40 a.m. – 12:20 p.m.

TBD

Kamalika Chaudhuri (UC San Diego)

12:20 – 2:30 p.m.

Lunch (on your own)

2:30 – 3:10 p.m.

Open-Source and Science in the Era of Foundation Models

Percy Liang (Stanford University)

3:10 – 3:50 p.m.

Domain Adaptation-- 20 years of theory chasing practice

Shai Ben-David (University of Waterloo)

3:50 – 4:10 p.m.

Break

4:10 – 5 p.m.

Panel Discussion

Rich Zemel (Columbia University),
Shai Ben-David (University of Waterloo and Vector Institute ),
Kamalika Chaudhuri (UC San Diego),
Percy Liang,
Hongseok Namkoong (Columbia University)