Few-Shot Learning of Compact Models via Task-Specific Meta Distillation

元学习(计算机科学) 计算机科学 软件部署 水准点(测量) 任务(项目管理) 人工智能 机器学习 适应(眼睛) 基线(sea) 弹丸 元建模 软件工程 光学 物理 地质学 经济 有机化学 化学 海洋学 管理 地理 大地测量学
作者
Yong Wu,Shekhor Chanda,Mehrdad Hosseinzadeh,Zhi Liu,Yan Wang
标识
DOI:10.1109/wacv56688.2023.00620
摘要

We consider a new problem of few-shot learning of com-pact models. Meta-learning is a popular approach for few-shot learning. Previous work in meta-learning typically assumes that the model architecture during meta-training is the same as the model architecture used for final deployment. In this paper, we challenge this basic assumption. For final deployment, we often need the model to be small. But small models usually do not have enough capacity to effectively adapt to new tasks. In the mean time, we often have access to the large dataset and extensive computing power during meta-training since meta-training is typically per-formed on a server. In this paper, we propose task-specific meta distillation that simultaneously learns two models in meta-learning: a large teacher model and a small student model. These two models are jointly learned during meta-training. Given a new task during meta-testing, the teacher model is first adapted to this task, then the adapted teacher model is used to guide the adaptation of the student model. The adapted student model is used for final deployment. We demonstrate the effectiveness of our approach in few-shot image classification using model-agnostic meta-learning (MAML). Our proposed method outperforms other alternatives on several benchmark datasets.
最长约 10秒,即可获得该文献文件

科研通智能强力驱动
Strongly Powered by AbleSci AI
更新
PDF的下载单位、IP信息已删除 (2025-6-4)

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
某某完成签到,获得积分10
刚刚
congenialboy发布了新的文献求助10
2秒前
孙燕应助ioio采纳,获得20
3秒前
Della发布了新的文献求助10
3秒前
Everything完成签到,获得积分10
4秒前
黄海发布了新的文献求助10
4秒前
5秒前
6秒前
6秒前
幸福大白发布了新的文献求助10
7秒前
7秒前
c2完成签到,获得积分10
7秒前
Hey发布了新的文献求助20
7秒前
7秒前
7秒前
科研力力完成签到 ,获得积分10
7秒前
8秒前
9秒前
9秒前
9秒前
幸福大白发布了新的文献求助10
9秒前
9秒前
c2发布了新的文献求助10
10秒前
lalala发布了新的文献求助10
10秒前
11秒前
殷勤的哈密瓜完成签到,获得积分10
11秒前
12秒前
阜睿发布了新的文献求助10
13秒前
紫竹魔笛发布了新的文献求助10
13秒前
高兴芯发布了新的文献求助10
14秒前
14秒前
展希希发布了新的文献求助10
15秒前
16秒前
李健应助liii采纳,获得10
16秒前
辛辛完成签到,获得积分10
18秒前
19秒前
科研通AI2S应助xiaoyuanbao1988采纳,获得10
19秒前
我是老大应助超男采纳,获得30
20秒前
范范发布了新的文献求助10
22秒前
Nn完成签到 ,获得积分10
22秒前
高分求助中
A new approach to the extrapolation of accelerated life test data 1000
ACSM’s Guidelines for Exercise Testing and Prescription, 12th edition 500
‘Unruly’ Children: Historical Fieldnotes and Learning Morality in a Taiwan Village (New Departures in Anthropology) 400
Indomethacinのヒトにおける経皮吸収 400
Phylogenetic study of the order Polydesmida (Myriapoda: Diplopoda) 370
基于可调谐半导体激光吸收光谱技术泄漏气体检测系统的研究 350
Robot-supported joining of reinforcement textiles with one-sided sewing heads 320
热门求助领域 (近24小时)
化学 材料科学 医学 生物 工程类 有机化学 生物化学 物理 内科学 纳米技术 计算机科学 化学工程 复合材料 遗传学 基因 物理化学 催化作用 冶金 细胞生物学 免疫学
热门帖子
关注 科研通微信公众号,转发送积分 3989510
求助须知:如何正确求助?哪些是违规求助? 3531756
关于积分的说明 11254536
捐赠科研通 3270255
什么是DOI,文献DOI怎么找? 1804947
邀请新用户注册赠送积分活动 882113
科研通“疑难数据库(出版商)”最低求助积分说明 809176