Recent advances of continual learning in computer vision: An overview

H Qu, H Rahmani, L Xu, B Williams, J Liu - arXiv preprint arXiv …, 2021 - arxiv.org
In contrast to batch learning where all training data is available at once, continual learning
represents a family of methods that accumulate knowledge and learn continuously with data …

Gdumb: A simple approach that questions our progress in continual learning

A Prabhu, PHS Torr, PK Dokania - … , Glasgow, UK, August 23–28, 2020 …, 2020 - Springer
We discuss a general formulation for the Continual Learning (CL) problem for classification—
a learning task where a stream provides samples to a learner and the goal of the learner …

“This is my unicorn, Fluffy”: Personalizing frozen vision-language representations

N Cohen, R Gal, EA Meirom, G Chechik… - European conference on …, 2022 - Springer
Abstract Large Vision & Language models pretrained on web-scale data provide
representations that are invaluable for numerous V &L problems. However, it is unclear how …

Wanderlust: Online continual object detection in the real world

J Wang, X Wang, Y Shang-Guan… - Proceedings of the …, 2021 - openaccess.thecvf.com
Online continual learning from data streams in dynamic environments is a critical direction in
the computer vision field. However, realistic benchmarks and fundamental studies in this line …

Learning to learn and remember super long multi-domain task sequence

Z Wang, L Shen, T Duan, D Zhan… - Proceedings of the …, 2022 - openaccess.thecvf.com
Catastrophic forgetting (CF) frequently occurs when learning with non-stationary data
distribution. The CF issue remains nearly unexplored and is more challenging when meta …

Meta-learning online adaptation of language models

N Hu, E Mitchell, CD Manning, C Finn - arXiv preprint arXiv:2305.15076, 2023 - arxiv.org
Large language models encode impressively broad world knowledge in their parameters.
However, the knowledge in static language models falls out of date, limiting the model's …

Strong Baselines for Parameter-Efficient Few-Shot Fine-Tuning

S Basu, S Hu, D Massiceti, S Feizi - … of the AAAI Conference on Artificial …, 2024 - ojs.aaai.org
Few-shot classification (FSC) entails learning novel classes given only a few examples per
class after a pre-training (or meta-training) phase on a set of base classes. Recent works …

When meta-learning meets online and continual learning: A survey

J Son, S Lee, G Kim - IEEE Transactions on Pattern Analysis …, 2024 - ieeexplore.ieee.org
Over the past decade, deep neural networks have demonstrated significant success using
the training scheme that involves mini-batch stochastic gradient descent on extensive …

Visual language navigation: A survey and open challenges

SM Park, YG Kim - Artificial Intelligence Review, 2023 - Springer
With the recent development of deep learning, AI models are widely used in various
domains. AI models show good performance for definite tasks such as image classification …

Memory-efficient semi-supervised continual learning: The world is its own replay buffer

J Smith, J Balloch, YC Hsu, Z Kira - 2021 International Joint …, 2021 - ieeexplore.ieee.org
Rehearsal is a critical component for class-incremental continual learning, yet it requires a
substantial memory budget. Our work investigates whether we can significantly reduce this …