Locality and compositionality in zero-shot learning
This work addresses the problem of improving generalization in ZSL for machine learning researchers, but it is incremental as it builds on existing concepts without introducing a new method.
The study investigated the roles of locality and compositionality in representation learning for Zero-Shot Learning (ZSL), finding that both properties are crucial for generalization, with experiments conducted without pre-training on external datasets like ImageNet.
In this work we study locality and compositionality in the context of learning representations for Zero Shot Learning (ZSL). In order to well-isolate the importance of these properties in learned representations, we impose the additional constraint that, differently from most recent work in ZSL, no pre-training on different datasets (e.g. ImageNet) is performed. The results of our experiments show how locality, in terms of small parts of the input, and compositionality, i.e. how well can the learned representations be expressed as a function of a smaller vocabulary, are both deeply related to generalization and motivate the focus on more local-aware models in future research directions for representation learning.