Timezone: »

Test-Time Training with Masked Autoencoders
Yossi Gandelsman · Yu Sun · Xinlei Chen · Alexei Efros

Wed Nov 30 09:00 AM -- 11:00 AM (PST) @ Hall J #915

Test-time training adapts to a new test distribution on the fly by optimizing a model for each test input using self-supervision.In this paper, we use masked autoencoders for this one-sample learning problem.Empirically, our simple method improves generalization on many visual benchmarks for distribution shifts.Theoretically, we characterize this improvement in terms of the bias-variance trade-off.

Author Information

Yossi Gandelsman (UC Berkeley)
Yu Sun (UC Berkeley)
Xinlei Chen (Facebook AI Research)
Alexei Efros (UC Berkeley)

More from the Same Authors