Open World Compositional Zero-Shot Learning
This work addresses the challenge of compositional zero-shot learning in a more realistic open world setting, which is incremental by extending existing methods to handle unfeasible compositions.
The paper tackles the problem of recognizing unseen state-object compositions in an open world setting, where many compositions are unfeasible, and shows that their method, using feasibility scores to mask or adjust cosine similarities, outperforms previous state-of-the-art approaches in this challenging scenario.
Compositional Zero-Shot learning (CZSL) requires to recognize state-object compositions unseen during training. In this work, instead of assuming prior knowledge about the unseen compositions, we operate in the open world setting, where the search space includes a large number of unseen compositions some of which might be unfeasible. In this setting, we start from the cosine similarity between visual features and compositional embeddings. After estimating the feasibility score of each composition, we use these scores to either directly mask the output space or as a margin for the cosine similarity between visual features and compositional embeddings during training. Our experiments on two standard CZSL benchmarks show that all the methods suffer severe performance degradation when applied in the open world setting. While our simple CZSL model achieves state-of-the-art performances in the closed world scenario, our feasibility scores boost the performance of our approach in the open world setting, clearly outperforming the previous state of the art.