Reproduce “Zero-Shot Learning Through Cross-Modal Transfer” 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
Richard Socher, Milind Ganjoo, Hamsa Sridhar, Osbert Bastani, Christopher D. Manning, Andrew Y. Ng
Socher et al. present a method for classifying images from categories without labeled examples while also predicting familiar classes. The system maps image features toward word vectors learned from unaligned text, then uses an outlier threshold to route images between seen and unseen classifiers. On CIFAR-10, at 70% seen-class accuracy, unseen-class accuracy is 15%–30%, depending on settings, with two classes held out.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.