Discrete Few-Shot Learning for Pan Privacy
This addresses privacy concerns for users of few-shot learning systems in resource-constrained settings, though it appears incremental as it builds on existing few-shot learning methods.
The paper tackles the problem of privacy risks in few-shot learning systems that store continuous embedding vectors by introducing discrete embedding vectors and a cryptographic protocol using one-way hash functions, achieving computational pan privacy without directly storing user embeddings.
In this paper we present the first baseline results for the task of few-shot learning of discrete embedding vectors for image recognition. Few-shot learning is a highly researched task, commonly leveraged by recognition systems that are resource constrained to train on a small number of images per class. Few-shot systems typically store a continuous embedding vector of each class, posing a risk to privacy where system breaches or insider threats are a concern. Using discrete embedding vectors, we devise a simple cryptographic protocol, which uses one-way hash functions in order to build recognition systems that do not store their users' embedding vectors directly, thus providing the guarantee of computational pan privacy in a practical and wide-spread setting.