Boosting Few-Shot Learning via Attentive Feature Regularization

Boosting(机器学习) 人工智能 正规化(语言学) 弹丸 计算机科学 特征(语言学) 模式识别(心理学) 机器学习 材料科学 哲学 语言学 冶金
作者
Xingyu Zhu,Shuo Wang,Jinda Lu,Yanbin Hao,Haifeng Liu,Xiangnan He
出处
期刊:Proceedings of the ... AAAI Conference on Artificial Intelligence [Association for the Advancement of Artificial Intelligence]
卷期号:38 (7): 7793-7801 被引量:18
标识
DOI:10.1609/aaai.v38i7.28614
摘要

Few-shot learning (FSL) based on manifold regularization aims to improve the recognition capacity of novel objects with limited training samples by mixing two samples from different categories with a blending factor. However, this mixing operation weakens the feature representation due to the linear interpolation and the overlooking of the importance of specific channels. To solve these issues, this paper proposes attentive feature regularization (AFR) which aims to improve the feature representativeness and discriminability. In our approach, we first calculate the relations between different categories of semantic labels to pick out the related features used for regularization. Then, we design two attention-based calculations at both the instance and channel levels. These calculations enable the regularization procedure to focus on two crucial aspects: the feature complementarity through adaptive interpolation in related categories and the emphasis on specific feature channels. Finally, we combine these regularization strategies to significantly improve the classifier performance. Empirical studies on several popular FSL benchmarks demonstrate the effectiveness of AFR, which improves the recognition accuracy of novel categories without the need to retrain any feature extractor, especially in the 1-shot setting. Furthermore, the proposed AFR can seamlessly integrate into other FSL methods to improve classification performance.

科研通智能强力驱动
Strongly Powered by AbleSci AI
科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
1秒前
称心青亦发布了新的文献求助10
2秒前
2秒前
2秒前
3秒前
4秒前
LN发布了新的文献求助30
5秒前
6秒前
6秒前
6秒前
科研通AI6.3的应助被XZM采纳,获得10
6秒前
7秒前
水天需发布了新的文献求助10
7秒前
banyingmm完成签到,获得积分10
8秒前
勤恳成威发布了新的文献求助10
8秒前
Nole的应助被王浩采纳,获得10
8秒前
9秒前
正直的夏青完成签到,获得积分20
9秒前
10秒前
xialuoke完成签到,获得积分10
10秒前
11秒前
111发布了新的文献求助10
12秒前
12秒前
banyingmm发布了新的文献求助10
13秒前
111完成签到,获得积分10
13秒前
我是老大的应助被贾永芳采纳,获得10
14秒前
儒雅一凤发布了新的文献求助10
14秒前
15秒前
煮香芋完成签到,获得积分10
16秒前
水天需发布了新的文献求助10
16秒前
16秒前
szhfendou发布了新的文献求助10
17秒前
Orange的应助被陈月亮采纳,获得10
17秒前
18秒前
123123完成签到 ,获得积分10
18秒前
18秒前
北漠狐发布了新的文献求助10
19秒前
趁风静完成签到 ,获得积分10
19秒前
vivi完成签到 ,获得积分10
20秒前
20秒前
高分求助中
(应助此贴封号)【重要!!请各用户(尤其是新用户)详细阅读】【科研通的精品贴汇总】 10000
拉伸高分子熔体中缠结行为的模拟和小角中子散射研究 1700
The World Atlas of Wine 9th edition 1000
奈特基础组织学:配套组织病理学(第4版)(Netter's Essential Histology With Correlated Histopathology) 800
八旗制度の研究 600
Improving the wear resistance of vehicle gear oil by oil soluble LaF3 nanoparticles and typical additives of gear oil 500
Boundaries and Bodhisattvas 500
热门求助领域 (近24小时)
化学 材料科学 医学 生物 纳米技术 计算机科学 工程类 有机化学 化学工程 生物化学 物理 复合材料 内科学 催化作用 光电子学 人工智能 细胞生物学 无机化学 物理化学 数学
热门帖子
关注 科研通微信公众号,转发送积分 8183923
求助须知:如何正确求助?哪些是违规求助? 9573132
关于积分的说明 21829461
捐赠科研通 7781024
什么是DOI,文献DOI怎么找? 3434150
关于科研通互助平台的介绍 2558333
邀请新用户注册赠送积分活动 2480648