Stochastic Prototype EmbeddingsDownload PDF

25 Sept 2019 (modified: 05 May 2023)ICLR 2020 Conference Blind SubmissionReaders: Everyone
Keywords: deep embeddings, stochastic embeddings, probabilistic embeddings, deep metric learning, few-shot learning
TL;DR: The paper proposes a probabilistic extension of Prototypical Networks that achieves superior few-shot, large-, and open-set classification performance, while gracefully handling label noise and out-of-distribution inputs.
Abstract: Supervised deep-embedding methods project inputs of a domain to a representational space in which same-class instances lie near one another and different-class instances lie far apart. We propose a probabilistic method that treats embeddings as random variables. Extending a state-of-the-art deterministic method, Prototypical Networks (Snell et al., 2017), our approach supposes the existence of a class prototype around which class instances are Gaussian distributed. The prototype posterior is a product distribution over labeled instances, and query instances are classified by marginalizing relative prototype proximity over embedding uncertainty. We describe an efficient sampler for approximate inference that allows us to train the model at roughly the same space and time cost as its deterministic sibling. Incorporating uncertainty improves performance on few-shot learning and gracefully handles label noise and out-of-distribution inputs. Compared to the state-of-the-art stochastic method, Hedged Instance Embeddings (Oh et al., 2019), we achieve superior large- and open-set classification accuracy. Our method also aligns class-discriminating features with the axes of the embedding space, yielding an interpretable, disentangled representation.
Code: https://drive.google.com/open?id=1QtShsWAHxfZRoxLG0Wcy4yVEfvGaqreG
Original Pdf: pdf
9 Replies

Loading