A survey on stability of learning with limited labelled data and its sensitivity to the effects of randomness

B Pecher, I Srba, M Bielikova - ACM Computing Surveys, 2024‏ - dl.acm.org
Learning with limited labelled data, such as prompting, in-context learning, fine-tuning, meta-
learning, or few-shot learning, aims to effectively train a model using only a small amount of …

Instance-based max-margin for practical few-shot recognition

M Fu, K Zhu - Proceedings of the IEEE/CVF Conference on …, 2024‏ - openaccess.thecvf.com
In order to mimic the human few-shot learning (FSL) ability better and to make FSL closer to
real-world applications this paper proposes a practical FSL (pFSL) setting. pFSL is based on …

NegCosIC: Negative cosine similarity-invariance-covariance regularization for few-shot learning

WH Liu, KM Lim, TS Ong, CP Lee - IEEE Access, 2024‏ - ieeexplore.ieee.org
Few-shot learning continues to pose a challenge as it is inherently difficult for visual
recognition models to generalize with limited labeled examples. When the training data is …

Task Attribute Distance for Few-Shot Learning: Theoretical Analysis and Applications

M Hu, H Chang, Z Guo, B Ma, S Shan… - arxiv preprint arxiv …, 2024‏ - arxiv.org
Few-shot learning (FSL) aims to learn novel tasks with very few labeled samples by
leveraging experience from\emph {related} training tasks. In this paper, we try to understand …

Towards Stabilized Few-Shot Object Detection with Less Forgetting via Sample Normalization

Y Ren, M Yang, Y Han, W Li - Sensors, 2024‏ - mdpi.com
Few-shot object detection is a challenging task aimed at recognizing novel classes and
localizing with limited labeled data. Although substantial achievements have been obtained …

[HTML][HTML] Few-Shot Learning in Wi-Fi-Based Indoor Positioning

F **e, SH Lam, M **e, C Wang - Biomimetics, 2024‏ - mdpi.com
This paper explores the use of few-shot learning in Wi-Fi-based indoor positioning, utilizing
convolutional neural networks (CNNs) combined with meta-learning techniques to enhance …

Building one-shot semi-supervised (BOSS) learning up to fully supervised performance

LN Smith, A Conovaloff - Frontiers in Artificial Intelligence, 2022‏ - frontiersin.org
Reaching the performance of fully supervised learning with unlabeled data and only
labeling one sample per class might be ideal for deep learning applications. We …

Few-Shot Object Detection Via Stabilized Meta-Learning Framework with Sample Normalization

Y Ren, H Jiang, M Yang - Available at SSRN 4576880‏ - papers.ssrn.com
Few-shot object detection has gained widespread attention due to its capability to transfer
knowledge from base classes to novel classes. However, existing methods suffer the …