Prototypical Networks for Few-shot Learning. Snell, J., Swersky, K., & Zemel, R. In Advances in Neural Information Processing Systems, volume 30, 2017. Curran Associates, Inc.. ![Prototypical Networks for Few-shot Learning [link]](https://bibbase.org/img/filetypes/link.svg "link")
Paper abstract bibtex We propose Prototypical Networks for the problem of few-shot classification, where a classifier must generalize to new classes not seen in the training set, given only a small number of examples of each new class. Prototypical Networks learn a metric space in which classification can be performed by computing distances to prototype representations of each class. Compared to recent approaches for few-shot learning, they reflect a simpler inductive bias that is beneficial in this limited-data regime, and achieve excellent results. We provide an analysis showing that some simple design decisions can yield substantial improvements over recent approaches involving complicated architectural choices and meta-learning. We further extend Prototypical Networks to zero-shot learning and achieve state-of-the-art results on the CU-Birds dataset.
@inproceedings{snell_prototypical_2017,
title = {Prototypical {Networks} for {Few}-shot {Learning}},
volume = {30},
url = {https://proceedings.neurips.cc/paper_files/paper/2017/hash/cb8da6767461f2812ae4290eac7cbc42-Abstract.html},
abstract = {We propose Prototypical Networks for the problem of few-shot classification, where a classifier must generalize to new classes not seen in the training set, given only a small number of examples of each new class. Prototypical Networks learn a metric space in which classification can be performed by computing distances to prototype representations of each class. Compared to recent approaches for few-shot learning, they reflect a simpler inductive bias that is beneficial in this limited-data regime, and achieve excellent results. We provide an analysis showing that some simple design decisions can yield substantial improvements over recent approaches involving complicated architectural choices and meta-learning. We further extend Prototypical Networks to zero-shot learning and achieve state-of-the-art results on the CU-Birds dataset.},
urldate = {2023-10-03},
booktitle = {Advances in {Neural} {Information} {Processing} {Systems}},
publisher = {Curran Associates, Inc.},
author = {Snell, Jake and Swersky, Kevin and Zemel, Richard},
year = {2017},
}
Downloads: 0
{"_id":"EE4HDvymrjsm4bNWL","bibbaseid":"snell-swersky-zemel-prototypicalnetworksforfewshotlearning-2017","authorIDs":[],"author_short":["Snell, J.","Swersky, K.","Zemel, R."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","title":"Prototypical Networks for Few-shot Learning","volume":"30","url":"https://proceedings.neurips.cc/paper_files/paper/2017/hash/cb8da6767461f2812ae4290eac7cbc42-Abstract.html","abstract":"We propose Prototypical Networks for the problem of few-shot classification, where a classifier must generalize to new classes not seen in the training set, given only a small number of examples of each new class. Prototypical Networks learn a metric space in which classification can be performed by computing distances to prototype representations of each class. Compared to recent approaches for few-shot learning, they reflect a simpler inductive bias that is beneficial in this limited-data regime, and achieve excellent results. We provide an analysis showing that some simple design decisions can yield substantial improvements over recent approaches involving complicated architectural choices and meta-learning. We further extend Prototypical Networks to zero-shot learning and achieve state-of-the-art results on the CU-Birds dataset.","urldate":"2023-10-03","booktitle":"Advances in Neural Information Processing Systems","publisher":"Curran Associates, Inc.","author":[{"propositions":[],"lastnames":["Snell"],"firstnames":["Jake"],"suffixes":[]},{"propositions":[],"lastnames":["Swersky"],"firstnames":["Kevin"],"suffixes":[]},{"propositions":[],"lastnames":["Zemel"],"firstnames":["Richard"],"suffixes":[]}],"year":"2017","bibtex":"@inproceedings{snell_prototypical_2017,\n\ttitle = {Prototypical {Networks} for {Few}-shot {Learning}},\n\tvolume = {30},\n\turl = {https://proceedings.neurips.cc/paper_files/paper/2017/hash/cb8da6767461f2812ae4290eac7cbc42-Abstract.html},\n\tabstract = {We propose Prototypical Networks for the problem of few-shot classification, where a classifier must generalize to new classes not seen in the training set, given only a small number of examples of each new class. Prototypical Networks learn a metric space in which classification can be performed by computing distances to prototype representations of each class. Compared to recent approaches for few-shot learning, they reflect a simpler inductive bias that is beneficial in this limited-data regime, and achieve excellent results. We provide an analysis showing that some simple design decisions can yield substantial improvements over recent approaches involving complicated architectural choices and meta-learning. We further extend Prototypical Networks to zero-shot learning and achieve state-of-the-art results on the CU-Birds dataset.},\n\turldate = {2023-10-03},\n\tbooktitle = {Advances in {Neural} {Information} {Processing} {Systems}},\n\tpublisher = {Curran Associates, Inc.},\n\tauthor = {Snell, Jake and Swersky, Kevin and Zemel, Richard},\n\tyear = {2017},\n}\n\n\n\n","author_short":["Snell, J.","Swersky, K.","Zemel, R."],"key":"snell_prototypical_2017","id":"snell_prototypical_2017","bibbaseid":"snell-swersky-zemel-prototypicalnetworksforfewshotlearning-2017","role":"author","urls":{"Paper":"https://proceedings.neurips.cc/paper_files/paper/2017/hash/cb8da6767461f2812ae4290eac7cbc42-Abstract.html"},"metadata":{"authorlinks":{}},"downloads":0,"html":""},"bibtype":"inproceedings","biburl":"https://bibbase.org/zotero/mh_lenguyen","creationDate":"2020-02-11T20:02:58.217Z","downloads":0,"keywords":[],"search_terms":["prototypical","networks","few","shot","learning","snell","swersky","zemel"],"title":"Prototypical Networks for Few-shot Learning","year":2017,"dataSources":["9ZC4Jom7ygp8s9git","iwKepCrWBps7ojhDx"]}