Saved in:
Bibliographic Details
Main Authors: Fifty, Christopher, Duan, Dennis, Junkins, Ronald G., Amid, Ehsan, Leskovec, Jure, Re, Christopher, Thrun, Sebastian
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2310.10971
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914728143486976
author Fifty, Christopher
Duan, Dennis
Junkins, Ronald G.
Amid, Ehsan
Leskovec, Jure
Re, Christopher
Thrun, Sebastian
author_facet Fifty, Christopher
Duan, Dennis
Junkins, Ronald G.
Amid, Ehsan
Leskovec, Jure
Re, Christopher
Thrun, Sebastian
contents Large Language Models like ChatGPT demonstrate a remarkable capacity to learn new concepts during inference without any fine-tuning. However, visual models trained to detect new objects during inference have been unable to replicate this ability, and instead either perform poorly or require meta-training and/or fine-tuning on similar objects. In this work, we propose a meta-learning algorithm that emulates Large Language Models by learning new visual concepts during inference without fine-tuning. Our approach leverages a frozen pre-trained feature extractor, and analogous to in-context learning, recasts visual meta-learning as sequence modeling over datapoints with known labels and a test datapoint with an unknown label. On 8 out of 11 meta-learning benchmarks, our approach -- without meta-training or fine-tuning -- exceeds or matches the state-of-the-art algorithm, P>M>F, which is meta-trained on these benchmarks. Our code is available at https://github.com/cfifty/CAML.
format Preprint
id arxiv_https___arxiv_org_abs_2310_10971
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Context-Aware Meta-Learning
Fifty, Christopher
Duan, Dennis
Junkins, Ronald G.
Amid, Ehsan
Leskovec, Jure
Re, Christopher
Thrun, Sebastian
Machine Learning
Computer Vision and Pattern Recognition
Large Language Models like ChatGPT demonstrate a remarkable capacity to learn new concepts during inference without any fine-tuning. However, visual models trained to detect new objects during inference have been unable to replicate this ability, and instead either perform poorly or require meta-training and/or fine-tuning on similar objects. In this work, we propose a meta-learning algorithm that emulates Large Language Models by learning new visual concepts during inference without fine-tuning. Our approach leverages a frozen pre-trained feature extractor, and analogous to in-context learning, recasts visual meta-learning as sequence modeling over datapoints with known labels and a test datapoint with an unknown label. On 8 out of 11 meta-learning benchmarks, our approach -- without meta-training or fine-tuning -- exceeds or matches the state-of-the-art algorithm, P>M>F, which is meta-trained on these benchmarks. Our code is available at https://github.com/cfifty/CAML.
title Context-Aware Meta-Learning
topic Machine Learning
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2310.10971