Reproduce “Revisiting Realistic Test-Time Training: Sequential Inference and Adaptation by Anchored Clustering” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Yongyi Su, Xun Xu, Kui Jia
TTAC enables sequential test-time training by adapting image classifiers as target examples arrive one at a time, without changing the source training objective or using later samples. It stores source-class feature statistics as anchors, clusters incoming target features, and aligns the clusters to those anchors. On CIFAR10-C, it reports 10.94% average error, versus 13.69% for protocol-adapted TTT++ under this protocol.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.