Meta Omnium: A Benchmark for General-Purpose Learning-to-Learn

Ondrej Bohdal, Yinbing Tian, Yongshuo Zong, Ruchika Chavhan, Da Li +3 more
5/12/2023
cs.CVcs.LGstat.ML

No official code found. Best implementation: edi-meta-learning/meta-omnium (25 stars, last commit 2023-06-19).

Abstract

Meta-learning and other approaches to few-shot learning are widely studied for image recognition, and are increasingly applied to other vision tasks such as pose estimation and dense prediction. This naturally raises the question of whether there is any few-shot meta-learning algorithm capable of generalizing across these diverse task types? To support the community in answering this question, we introduce Meta Omnium, a dataset-of-datasets spanning multiple vision tasks including recognition, keypoint localization, semantic segmentation and regression. We experiment with popular few-shot meta-learning baselines and analyze their ability to generalize across tasks and to transfer knowledge between them. Meta Omnium enables meta-learning researchers to evaluate model generalization to a much wider array of tasks than previously possible, and provides a single framework for evaluating meta-learners across a wide suite of vision applications in a consistent manner.

View on arXivView PDF

Code Implementations(1)

edi-meta-learning/meta-omniumCommunity100%

Implementation of "Meta Omnium: A Benchmark for General-Purpose Learning-to-Learn"

253Shell, PythonMar 27, 20233 years ago
deep-learningfew-shot-learning

Similar Papers

Meta-learning approaches for few-shot learning: A survey of recent advances

86% match2023

Meta-Task: A Method-Agnostic Framework for Learning to Regularize in Few-Shot Learning

85% match2024

Rethinking Few-Shot Image Classification: a Good Embedding Is All You Need?

85% match1 repo3752020

Meta Learning to Bridge Vision and Language Models for Multimodal Few-Shot Learning

84% match1 repo612023

Multimodal CLIP Inference for Meta-Few-Shot Image Classification

83% match2024

Bayesian Meta-Learning for the Few-Shot Setting via Deep Kernels

83% match1 repo2082019

Cite this paper

@article{bohdal2023meta,
  title  = {Meta Omnium: A Benchmark for General-Purpose Learning-to-Learn},
  author = {Ondrej Bohdal and Yinbing Tian and Yongshuo Zong and Ruchika Chavhan and Da Li and Henry Gouk and Li Guo and Timothy Hospedales},
  year   = {2023},
  eprint = {2305.07625},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2305.07625v1},
  journal = {CVPR 2023 2023}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2305.07625.svg)](https://codeofpaper.com/paper/2305.07625)

Discussion