Filters








12,321 Hits in 4.8 sec

Prior-Knowledge and Attention-based Meta-Learning for Few-Shot Learning [article]

Yunxiao Qin, Weiguo Zhang, Chenxu Zhao, Zezheng Wang, Xiangyu Zhu, Guojun Qi, Jingping Shi, Zhen Lei
2020 arXiv   pre-print
developments to introduce attention mechanism and prior-knowledge for meta-learning.  ...  Compared with existing meta-learning approaches that pay little attention to prior-knowledge and vision attention, our approach alleviates the meta-learner's few-shot cognition burden.  ...  In our paper, we improve meta-learning for few-shot learning by introducing prior-knowledge and attention mechanism to meta-learning.  ... 
arXiv:1812.04955v5 fatcat:na3ro6wfvre4xmaazrzmoglpge

Task Attended Meta-Learning for Few-Shot Learning [article]

Aroof Aimen, Sahil Sidheekh, Narayanan C. Krishnan
2021 arXiv   pre-print
Meta-learning (ML) has emerged as a promising direction in learning models under constrained resource settings like few-shot learning.  ...  The former approaches leverage the knowledge from a batch of tasks to learn an optimal prior. In this work, we study the importance of a batch for ML.  ...  Acknowledgements The support and the resources provided by 'PARAM Shivay Facility' under the National Supercomputing Mission, Government of India at the Indian Institute of Technology, Varanasi and under  ... 
arXiv:2106.10642v1 fatcat:qa5lae7id5fefkd23vzvtvp7ny

A Concise Review of Recent Few-shot Meta-learning Methods [article]

Xiaoxu Li and Zhuo Sun and Jing-Hao Xue and Zhanyu Ma
2020 arXiv   pre-print
Few-shot meta-learning has been recently reviving with expectations to mimic humanity's fast adaption to new concepts based on prior knowledge.  ...  We conclude this review with some vital current challenges and future prospects in few-shot meta-learning.  ...  In the cases of few-shot meta-learning, a meta-learner is trained to learn some prior or shared knowledge from  , and then modified on tasks on  .  ... 
arXiv:2005.10953v1 fatcat:v54jrpktazf3bfx4kqos4ls27y

LGM-Net: Learning to Generate Matching Networks for Few-Shot Learning [article]

Huaiyu Li, Weiming Dong, Xing Mei, Chongyang Ma, Feiyue Huang, Bao-Gang Hu
2019 arXiv   pre-print
In this work, we propose a novel meta-learning approach for few-shot classification, which learns transferable prior knowledge across tasks and directly produces network parameters for similar unseen tasks  ...  show that transferable prior knowledge is learned by the MetaNet module via mapping training data to functional weights.  ...  Acknowledgements This work was supported by National Key R&D Program of China under no. 2018YFC0807500, National Natural Science Foundation of China under nos. 61832016, 61720106006 and 61672520, as well  ... 
arXiv:1905.06331v1 fatcat:6jnu4cvw65e23mdgytkpu3dfuq

XtarNet: Learning to Extract Task-Adaptive Representation for Incremental Few-Shot Learning [article]

Sung Whan Yoon, Do-Yeon Kim, Jun Seo, Jaekyun Moon
2020 arXiv   pre-print
Learning novel concepts while preserving prior knowledge is a long-standing challenge in machine learning.  ...  We propose XtarNet, which learns to extract task-adaptive representation (TAR) for facilitating incremental few-shot learning.  ...  Acknowledgements This work was supported by IITP funds from MSIT of Korea (No. 2016-0-00563, No. 2020-0-00626 and No. 2020 for KAIST and AI Graduate School Program at UNIST.  ... 
arXiv:2003.08561v2 fatcat:4eup33cncfdivae37x55wvtk6q

COMPAS: Representation Learning with Compositional Part Sharing for Few-Shot Classification [article]

Ju He, Adam Kortylewski, Alan Yuille
2021 arXiv   pre-print
Few-shot image classification consists of two consecutive learning processes: 1) In the meta-learning stage, the model acquires a knowledge base from a set of training classes. 2) During meta-testing,  ...  During meta-testing, the representation of unseen classes is learned using the part representations and the part activation maps from the knowledge base.  ...  Few-Shot Classification Few-shot image classification consists of two consecutive learning processes: 1) In the meta-learning stage, the model acquires a knowledge base from a set of training classes.  ... 
arXiv:2101.11878v2 fatcat:gker3nb255acpjss7rdkvdte4y

Few-Shot Image Classification: Current Status and Research Trends

Ying Liu, Hengchang Zhang, Weidong Zhang, Guojun Lu, Qi Tian, Nam Ling
2022 Electronics  
Meta-learning based methods employ past prior knowledge to guide the learning of new tasks. Data augmentation based methods expand the amount of sample data with auxiliary information.  ...  Transfer learning based methods transfer useful prior knowledge from the source domain to the target domain.  ...  Transfer learning can transfer the useful prior knowledge from the source domain to the target domain, which is conducive to few-shot learning; meta-learning employs the prior knowledge learned from a  ... 
doi:10.3390/electronics11111752 fatcat:tsc53f5c5fe6fdswsm3ljnvn3e

Challenge Closed-book Science Exam: A Meta-learning Based Question Answering System [article]

Xinyue Zheng, Peng Wang, Qigang Wang, Zhongchao Shi
2020 arXiv   pre-print
knowledge bases.  ...  Specifically, our method based on meta-learning method and large language model BERT, which can efficiently solve science problems by learning from related example questions without relying on external  ...  K b do 5: Fig. 3 . 3 Attention-head view for few-shot learning model, for the input text Which substance is magnetic and conducts heat?  ... 
arXiv:2004.12303v1 fatcat:5xzmebvh2vbtpi2z7ogfkvmhau

Learning to Learn Kernels with Variational Random Features [article]

Xiantong Zhen, Haoliang Sun, Yingjun Du, Jun Xu, Yilong Yin, Ling Shao, Cees Snoek
2020 arXiv   pre-print
In this work, we introduce kernels with random Fourier features in the meta-learning framework to leverage their strong few-shot learning ability.  ...  Experimental results on a variety of few-shot regression and classification tasks demonstrate that MetaVRF delivers much better, or at least competitive, performance compared to existing meta-learning  ...  Method We first describe the base-learner based on the kernel ridge regression in meta-learning for few-shot learning, and then introduce kernel learning with random features, based on which our meta variational  ... 
arXiv:2006.06707v2 fatcat:cnw4cqcj4jcddculybknqgmqcq

Meta-SE: A meta-learning framework for few-shot speech enhancement

Weili Zhou, Mingliang Lu, Ruijie Ji
2021 IEEE Access  
PRELIMINARY FOR META-LEARNING Meta-learning has become the research focus of few-shot learning due to its capability of quickly process new tasks with few samples by the prior meta-knowledge.  ...  new tasks with few samples by the prior meta-knowledge.  ... 
doi:10.1109/access.2021.3066609 fatcat:kjdirvjkwbgmbayngkx7dreq5q

Toward Multimodal Model-Agnostic Meta-Learning [article]

Risto Vuorio, Shao-Hua Sun, Hexiang Hu, Joseph J. Lim
2018 arXiv   pre-print
Gradient-based meta-learners such as MAML are able to learn a meta-prior from similar tasks to adapt to novel tasks from the same distribution with few gradient updates.  ...  We evaluate the proposed model on a diverse set of problems including regression, few-shot image classification, and reinforcement learning.  ...  Optimization-based meta-learning methods offer learnable learning rules and optimization algorithms [21, 2, 19, 1, 8] , metric-based meta learners [11, 31, 26, 25, 27] address few-shot classification  ... 
arXiv:1812.07172v1 fatcat:f76cyhyu2jc2zmoen4pmqs2luq

Few-shot learning for medical text: A systematic review [article]

Yao Ge, Yuting Guo, Yuan-Chi Yang, Mohammed Ali Al-Garadi, Abeed Sarker
2022 arXiv   pre-print
Objective: Few-shot learning (FSL) methods require small numbers of labeled instances for training.  ...  Common methods included FSL with attention mechanisms (12/31; 39%), prototypical networks (8/31; 26%), and meta-learning (6/31; 19%).  ...  Architectures of three popular few-shot learning methodologies. (a) Meta-learning: each task mimics the few-shot scenario, and can be completely non-overlapping.  ... 
arXiv:2204.14081v1 fatcat:ageqcud25fh3xeuctrgeqytmhe

TAE-Net: Task-Adaptive Embedding Network for Few-Shot Remote Sensing Scene Classification

Wendong Huang, Zhengwu Yuan, Aixia Yang, Chan Tang, Xiaobo Luo
2021 Remote Sensing  
In this study, few-shot learning is used for remote sensing scene classification tasks. The goal of few-shot learning is to recognize unseen scene categories given extremely limited labeled samples.  ...  The experimental results illustrate that our proposed TAE-Net achieves new state-of-the-art performance for few-shot remote sensing scene classification.  ...  Different with standard supervised learning, there are extremely few labeled samples available in few-shot learning, which makes it difficult for the model to learn enough prior knowledge.  ... 
doi:10.3390/rs14010111 fatcat:6qjuedqtabhzro67e7tsikbweq

Task-wise attention guided part complementary learning for few-shot image classification

Gong Cheng, Ruimin Li, Chunbo Lang, Junwei Han
2021 Science China Information Sciences  
A general framework to tackle the problem of few-shot learning is meta-learning, which aims to train a well-generalized meta-learner (or backbone network) to learn a base-learner for each future task with  ...  In fact, the learning of base-learners acting with each specific task is also significantly crucial for few-shot learning.  ...  By leveraging the base datasets with sufficient samples, a capable meta-learner can be developed in the meta-training phase, and then provides prior knowledge (also known as meta-knowledge) for base-learners  ... 
doi:10.1007/s11432-020-3156-7 fatcat:hl6evdxcqraflnkpbvjtzrqoju

Target unbiased meta-learning for graph classification

Ming Li, Shuo Zhu, Chunxu Li, Wencang Zhao
2021 Journal of Computational Design and Engineering  
Even though numerous works focus on the few-shot learning issue by combining meta-learning, there are still limits to traditional graph classification problems.  ...  In order to overcome the aforementioned issues, an effective strategy with training an unbiased meta-learning algorithm was developed in this paper, which sorted out problems of target preference and few-shot  ...  Bias: Most meta-learning tasks coexist with multiple tasks, and they have a certain memory ability for prior knowledge.  ... 
doi:10.1093/jcde/qwab050 fatcat:oydtndulcnhxboq4fjfp674rja
« Previous Showing results 1 — 15 out of 12,321 results