Transfer Learning

We explain how visual attribute learning may benefit from multi-task/transfer learning approaches that carefully select what to share among whom. We show that such methods can combat noisy correlations among attributes, and variations among attribute appearances across object categories.

While learning multiple attributes with possibly noisy correlations in the training set, it helps to employ a multi-task learning approach that tries to learn classifiers that rely on different parts of the input feature space.

Zero-shot recognition systems often rely on visual attribute classifiers, which may be noisy. However, this noise is systematic, and if modeled correctly and used together with our proposed approach, zero-shot learning outcomes can be significantly improved.