Practical Aspects of Zero-Shot Learning
This work addresses the practical challenge of selecting effective zero-shot learning methods for applications where labeled training data is unavailable, though it is incremental in nature.
The paper tackles the lack of a dominant algorithm in zero-shot learning by comparing state-of-the-art methods on benchmark datasets and proposing meta-classifiers that combine their best aspects, with experimental results showing improved performance.
One of important areas of machine learning research is zero-shot learning. It is applied when properly labeled training data set is not available. A number of zero-shot algorithms have been proposed and experimented with. However, none of them seems to be the "overall winner". In situations like this, it may be possible to develop a meta-classifier that would combine "best aspects" of individual classifiers and outperform all of them. In this context, the goal of this contribution is twofold. First, multiple state-of-the-art zero-shot learning methods are compared for standard benchmark datasets. Second, multiple meta-classifiers are suggested and experimentally compared (for the same datasets).