{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/gaussian-prototypical-networks-for-few-shot","title":"Gaussian Prototypical Networks for Few-Shot Learning on Omniglot","arxiv_id":"1708.02735","date":"2017-08-09","proceeding":"ICLR 2018 1","authors":["Stanislav Fort"],"abstract":"We propose a novel architecture for $k$-shot classification on the Omniglot\ndataset. Building on prototypical networks, we extend their architecture to\nwhat we call Gaussian prototypical networks. Prototypical networks learn a map\nbetween images and embedding vectors, and use their clustering for\nclassification. In our model, a part of the encoder output is interpreted as a\nconfidence region estimate about the embedding point, and expressed as a\nGaussian covariance matrix. Our network then constructs a direction and class\ndependent distance metric on the embedding space, using uncertainties of\nindividual data points as weights. We show that Gaussian prototypical networks\nare a preferred architecture over vanilla prototypical networks with an\nequivalent number of parameters. We report state-of-the-art performance in\n1-shot and 5-shot classification both in 5-way and 20-way regime (for 5-shot\n5-way, we are comparable to previous state-of-the-art) on the Omniglot dataset.\nWe explore artificially down-sampling a fraction of images in the training set,\nwhich improves our performance even further. We therefore hypothesize that\nGaussian prototypical networks might perform better in less homogeneous,\nnoisier datasets, which are commonplace in real world applications.","url_abs":"http://arxiv.org/abs/1708.02735v1","url_pdf":"http://arxiv.org/pdf/1708.02735v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"gaussian-prototypical-networks-for-few-shot","repo_url":"https://github.com/stanislavfort/gaussian-prototypical-networks","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"classification-1","task_name":"Classification"},{"task_slug":"clustering","task_name":"Clustering"},{"task_slug":"few-shot-learning","task_name":"Few-Shot Learning"},{"task_slug":"classification","task_name":"General Classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1708.02735","atlas_url":"https://app.syntology.ai/?focus=1708.02735","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}