Goto

Collaborating Authors

 Supervised Learning



Geometry-Aware Adaptation for Pretrained Models

Neural Information Processing Systems

Machine learning models--including prominent zero-shot models--are often trained on datasets whose labels are only a small proportion of a larger label space. Such spaces are commonly equipped with a metric that relates the labels via distances between them.








LiftingWeakSupervisionToStructuredPrediction

Neural Information Processing Systems

For labels taking values in a finite metric space, we introduce techniques new to weak supervision based on pseudo-Euclidean embeddings andtensor decompositions, providing anearly-consistent noise rate estimator.


LiftingWeakSupervisionToStructuredPrediction

Neural Information Processing Systems

For labels taking values in a finite metric space, we introduce techniques new to weak supervision based on pseudo-Euclidean embeddings andtensor decompositions, providing anearly-consistent noise rate estimator.