HyperShot: Few-Shot Learning by Kernel HyperNetworks

Marcin Sendera, Marcin Przewięźlikowski, Konrad Karanowski, Maciej Zięba, Jacek Tabor, Przemysław Spurek; Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2023, pp. 2469-2478


Few-shot models aim at making predictions using a minimal number of labeled examples from a given task. The main challenge in this area is the one-shot setting where only one element represents each class. We propose HyperShot - the fusion of kernels and hypernetwork paradigm. Compared to reference approaches that apply a gradient-based adjustment of the parameters, our model aims to switch the classification module parameters depending on the task's embedding. In practice, we utilize a hypernetwork, which takes the aggregated information from support data and returns the classifier's parameters handcrafted for the considered problem. Moreover, we introduce the kernel-based representation of the support examples delivered to hypernetwork to create the parameters of the classification module. Consequently, we rely on relations between embeddings of the support examples instead of direct feature values provided by the backbone models. Thanks to this approach, our model can adapt to highly different tasks.

Related Material

[pdf] [supp]
@InProceedings{Sendera_2023_WACV, author = {Sendera, Marcin and Przewi\k{e}\'zlikowski, Marcin and Karanowski, Konrad and Zi\k{e}ba, Maciej and Tabor, Jacek and Spurek, Przemys{\l}aw}, title = {HyperShot: Few-Shot Learning by Kernel HyperNetworks}, booktitle = {Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)}, month = {January}, year = {2023}, pages = {2469-2478} }