dc.contributor.author Ayyad, Ahmed dc.contributor.author Navab, Nassir dc.contributor.author Elhoseiny, Mohamed dc.contributor.author Albarqouni, Shadi dc.date.accessioned 2020-04-27T01:43:07Z dc.date.available 2020-04-27T01:43:07Z dc.date.issued 2020-02-10 dc.date.submitted 2019-03-06 dc.identifier.uri http://hdl.handle.net/10754/662647 dc.description.abstract Recent progress has shown that few-shot learning can be improved with access to unlabelled data, known as semi-supervised few-shot learning(SS-FSL). We introduce an SS-FSL approach, dubbed as Prototypical Random Walk Networks(PRWN), built on top of Prototypical Networks (PN). We develop a random walk semi-supervised loss that enables the network to learn representations that are compact and well-separated. Our work is related to the very recent development on graph-based approaches for few-shot learning. However, we show that compact and well-separated class representations can be achieved by modeling our prototypical random walk notion without needing additional graph-NN parameters or requiring a transductive setting where collective test set is provided. Our model outperforms prior art in most benchmarks with significant improvements in some cases. For example, in a mini-Imagenet 5-shot classification task, we obtain 69.65$\%$ accuracy to the 64.59$\%$ state-of-the-art. Our model, trained with 40$\%$ of the data as labelled, compares competitively against fully supervised prototypical networks, trained on 100$\%$ of the labels, even outperforming it in the 1-shot mini-Imagenet case with 50.89$\%$ to 49.4$\%$ accuracy. We also show that our model is resistant to distractors, unlabeled data that does not belong to any of the training classes, and hence reflecting robustness to labelled/unlabelled class distribution mismatch. We also performed a challenging discriminative power test, showing a relative improvement on top of the baseline of $\approx$14\% on 20 classes on mini-Imagenet and $\approx$60\% on 800 classes on Omniglot. Code will be made available. dc.publisher arXiv dc.relation.url https://arxiv.org/pdf/1903.02164 dc.rights Archived with thanks to arXiv dc.title Semi-Supervised Few-Shot Learning with Prototypical Random Walks dc.type Preprint dc.contributor.department Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division dc.eprint.version Pre-print dc.contributor.institution Dept. of Informatics, TU Munich dc.contributor.institution Computer Aided Medical Procedures, Johns Hopkins University dc.identifier.arxivid 1903.02164 kaust.person Elhoseiny, Mohamed refterms.dateFOA 2020-04-27T01:43:44Z
﻿

### Files in this item

Name:
Preprintfile1.pdf
Size:
816.4Kb
Format:
PDF
Description:
Pre-print