Download free for 30 days
Sign in
Upload
Language (EN)
Support
Business
Mobile
Social Media
Marketing
Technology
Art & Photos
Career
Design
Education
Presentations & Public Speaking
Government & Nonprofit
Healthcare
Internet
Law
Leadership & Management
Automotive
Engineering
Software
Recruiting & HR
Retail
Sales
Services
Science
Small Business & Entrepreneurship
Food
Environment
Economy & Finance
Data & Analytics
Investor Relations
Sports
Spiritual
News & Politics
Travel
Self Improvement
Real Estate
Entertainment & Humor
Health & Medicine
Devices & Hardware
Lifestyle
Change Language
Language
English
Español
Português
Français
Deutsche
Cancel
Save
EN
Uploaded by
Deep Learning JP
PPTX, PDF
985 views
[DL輪読会]Exploiting Cyclic Symmetry in Convolutional Neural Networks
2017/2/28 Deep Learning JP: http://deeplearning.jp/seminar-2/
Technology
◦
Related topics:
Deep Learning
•
Read more
0
Save
Share
Embed
Embed presentation
Download
Download to read offline
1
/ 18
2
/ 18
3
/ 18
4
/ 18
5
/ 18
6
/ 18
7
/ 18
8
/ 18
9
/ 18
10
/ 18
11
/ 18
12
/ 18
13
/ 18
14
/ 18
15
/ 18
16
/ 18
17
/ 18
18
/ 18
More Related Content
PDF
[DL輪読会]Combining Fully Convolutional and Recurrent Neural Networks for 3D Bio...
by
Deep Learning JP
PDF
[DL輪読会]Wasserstein GAN/Towards Principled Methods for Training Generative Adv...
by
Deep Learning JP
PPTX
Deep sets
by
Tomohiro Takahashi
PPTX
[DL輪読会]Unsupervised Learning of Probably Symmetric Deformable 3D Objects from...
by
Deep Learning JP
PPTX
PRML 5.5.6-5.6 畳み込みネットワーク(CNN)・ソフト重み共有・混合密度ネットワーク
by
KokiTakamiya
PDF
cvpaper.challenge in CVPR2015 (PRMU2015年12月)
by
cvpaper. challenge
PPTX
Unsupervised Learning of Probably Symmetric Deformable 3D Objects from Images...
by
Kento Doi
PPTX
Image net classification with Deep Convolutional Neural Networks
by
Shingo Horiuchi
[DL輪読会]Combining Fully Convolutional and Recurrent Neural Networks for 3D Bio...
by
Deep Learning JP
[DL輪読会]Wasserstein GAN/Towards Principled Methods for Training Generative Adv...
by
Deep Learning JP
Deep sets
by
Tomohiro Takahashi
[DL輪読会]Unsupervised Learning of Probably Symmetric Deformable 3D Objects from...
by
Deep Learning JP
PRML 5.5.6-5.6 畳み込みネットワーク(CNN)・ソフト重み共有・混合密度ネットワーク
by
KokiTakamiya
cvpaper.challenge in CVPR2015 (PRMU2015年12月)
by
cvpaper. challenge
Unsupervised Learning of Probably Symmetric Deformable 3D Objects from Images...
by
Kento Doi
Image net classification with Deep Convolutional Neural Networks
by
Shingo Horiuchi
Similar to [DL輪読会]Exploiting Cyclic Symmetry in Convolutional Neural Networks
PDF
Learning Spatial Common Sense with Geometry-Aware Recurrent Networks
by
Kento Doi
PPTX
Cvim saisentan-6-4-tomoaki
by
tomoaki0705
PDF
20130925.deeplearning
by
Hayaru SHOUNO
PPTX
CVPR 2017 報告
by
Yu Nishimura
PDF
【2015.07】(2/2)cvpaper.challenge@CVPR2015
by
cvpaper. challenge
PDF
Deep SimNets
by
Fujimoto Keisuke
PDF
2019/5/24 Chainer familyで始める深層学習 ハンズオンの部
by
belltailjp
PPTX
PredCNN: Predictive Learning with Cascade Convolutions
by
harmonylab
PPTX
Eccv2018 report day4
by
Atsushi Hashimoto
PDF
(文献紹介)深層学習による動被写体ロバストなカメラの動き推定
by
Morpho, Inc.
PPTX
深層学習とTensorFlow入門
by
tak9029
PDF
NVIDIA Seminar ディープラーニングによる画像認識と応用事例
by
Takayoshi Yamashita
PPT
Deep Learningの技術と未来
by
Seiya Tokui
PDF
効率的学習 / Efficient Training(メタサーベイ)
by
cvpaper. challenge
PDF
論文読み会(DeMoN;CVPR2017)
by
Masaya Kaneko
PDF
dl-with-python01_handout
by
Shin Asakawa
PDF
Convolutional Neural Networks のトレンド @WBAFLカジュアルトーク#2
by
Daiki Shimada
PDF
Deep Learningの基礎と応用
by
Seiya Tokui
PDF
[DL輪読会]EdgeConnect: Generative Image Inpainting with Adversarial Edge Learning
by
Deep Learning JP
PDF
3DFeat-Net
by
Takuya Minagawa
Learning Spatial Common Sense with Geometry-Aware Recurrent Networks
by
Kento Doi
Cvim saisentan-6-4-tomoaki
by
tomoaki0705
20130925.deeplearning
by
Hayaru SHOUNO
CVPR 2017 報告
by
Yu Nishimura
【2015.07】(2/2)cvpaper.challenge@CVPR2015
by
cvpaper. challenge
Deep SimNets
by
Fujimoto Keisuke
2019/5/24 Chainer familyで始める深層学習 ハンズオンの部
by
belltailjp
PredCNN: Predictive Learning with Cascade Convolutions
by
harmonylab
Eccv2018 report day4
by
Atsushi Hashimoto
(文献紹介)深層学習による動被写体ロバストなカメラの動き推定
by
Morpho, Inc.
深層学習とTensorFlow入門
by
tak9029
NVIDIA Seminar ディープラーニングによる画像認識と応用事例
by
Takayoshi Yamashita
Deep Learningの技術と未来
by
Seiya Tokui
効率的学習 / Efficient Training(メタサーベイ)
by
cvpaper. challenge
論文読み会(DeMoN;CVPR2017)
by
Masaya Kaneko
dl-with-python01_handout
by
Shin Asakawa
Convolutional Neural Networks のトレンド @WBAFLカジュアルトーク#2
by
Daiki Shimada
Deep Learningの基礎と応用
by
Seiya Tokui
[DL輪読会]EdgeConnect: Generative Image Inpainting with Adversarial Edge Learning
by
Deep Learning JP
3DFeat-Net
by
Takuya Minagawa
More from Deep Learning JP
PPTX
【DL輪読会】AdaptDiffuser: Diffusion Models as Adaptive Self-evolving Planners
by
Deep Learning JP
PPTX
【DL輪読会】事前学習用データセットについて
by
Deep Learning JP
PPTX
【DL輪読会】 "Learning to render novel views from wide-baseline stereo pairs." CVP...
by
Deep Learning JP
PPTX
【DL輪読会】Zero-Shot Dual-Lens Super-Resolution
by
Deep Learning JP
PPTX
【DL輪読会】BloombergGPT: A Large Language Model for Finance arxiv
by
Deep Learning JP
PPTX
【DL輪読会】マルチモーダル LLM
by
Deep Learning JP
PDF
【 DL輪読会】ToolLLM: Facilitating Large Language Models to Master 16000+ Real-wo...
by
Deep Learning JP
PPTX
【DL輪読会】AnyLoc: Towards Universal Visual Place Recognition
by
Deep Learning JP
PDF
【DL輪読会】Can Neural Network Memorization Be Localized?
by
Deep Learning JP
PPTX
【DL輪読会】Hopfield network 関連研究について
by
Deep Learning JP
PPTX
【DL輪読会】SimPer: Simple self-supervised learning of periodic targets( ICLR 2023 )
by
Deep Learning JP
PDF
【DL輪読会】RLCD: Reinforcement Learning from Contrast Distillation for Language M...
by
Deep Learning JP
PDF
【DL輪読会】"Secrets of RLHF in Large Language Models Part I: PPO"
by
Deep Learning JP
PPTX
【DL輪読会】"Language Instructed Reinforcement Learning for Human-AI Coordination "
by
Deep Learning JP
PPTX
【DL輪読会】Llama 2: Open Foundation and Fine-Tuned Chat Models
by
Deep Learning JP
PDF
【DL輪読会】"Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware"
by
Deep Learning JP
PPTX
【DL輪読会】Parameter is Not All You Need:Starting from Non-Parametric Networks fo...
by
Deep Learning JP
PDF
【DL輪読会】Drag Your GAN: Interactive Point-based Manipulation on the Generative ...
by
Deep Learning JP
PDF
【DL輪読会】Self-Supervised Learning from Images with a Joint-Embedding Predictive...
by
Deep Learning JP
PPTX
【DL輪読会】Towards Understanding Ensemble, Knowledge Distillation and Self-Distil...
by
Deep Learning JP
【DL輪読会】AdaptDiffuser: Diffusion Models as Adaptive Self-evolving Planners
by
Deep Learning JP
【DL輪読会】事前学習用データセットについて
by
Deep Learning JP
【DL輪読会】 "Learning to render novel views from wide-baseline stereo pairs." CVP...
by
Deep Learning JP
【DL輪読会】Zero-Shot Dual-Lens Super-Resolution
by
Deep Learning JP
【DL輪読会】BloombergGPT: A Large Language Model for Finance arxiv
by
Deep Learning JP
【DL輪読会】マルチモーダル LLM
by
Deep Learning JP
【 DL輪読会】ToolLLM: Facilitating Large Language Models to Master 16000+ Real-wo...
by
Deep Learning JP
【DL輪読会】AnyLoc: Towards Universal Visual Place Recognition
by
Deep Learning JP
【DL輪読会】Can Neural Network Memorization Be Localized?
by
Deep Learning JP
【DL輪読会】Hopfield network 関連研究について
by
Deep Learning JP
【DL輪読会】SimPer: Simple self-supervised learning of periodic targets( ICLR 2023 )
by
Deep Learning JP
【DL輪読会】RLCD: Reinforcement Learning from Contrast Distillation for Language M...
by
Deep Learning JP
【DL輪読会】"Secrets of RLHF in Large Language Models Part I: PPO"
by
Deep Learning JP
【DL輪読会】"Language Instructed Reinforcement Learning for Human-AI Coordination "
by
Deep Learning JP
【DL輪読会】Llama 2: Open Foundation and Fine-Tuned Chat Models
by
Deep Learning JP
【DL輪読会】"Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware"
by
Deep Learning JP
【DL輪読会】Parameter is Not All You Need:Starting from Non-Parametric Networks fo...
by
Deep Learning JP
【DL輪読会】Drag Your GAN: Interactive Point-based Manipulation on the Generative ...
by
Deep Learning JP
【DL輪読会】Self-Supervised Learning from Images with a Joint-Embedding Predictive...
by
Deep Learning JP
【DL輪読会】Towards Understanding Ensemble, Knowledge Distillation and Self-Distil...
by
Deep Learning JP
[DL輪読会]Exploiting Cyclic Symmetry in Convolutional Neural Networks
1.
2017/2/16 DL輪読会 医学系研究科 D2 山口亮平
2.
Exploiting Cyclic Symmetry
in Convolutional Neural Networks • Deep Mindの論文 • Sander Dileman(ゲント大学の博士課程→google)。博士 課程のとき、Kaggle2015のプランクトンコンテストで 優勝。 • ICML16で発表、引用数12(2017/2/17現在) • http://benanne.github.io/2015/03/17/plankton.html • 回転に対するロバスト性をネットワーク内部で再現し た論文。
3.
Motivation <CNN・・・平行移動に対して強い> • Conv層・・・平行移動に対してequivariant • Pooling層・・・平行移動に対してinvariant → ①rotationに対してもinvariantを保ちたい。 ネットワークにそのinvariance/equivariantな性質を組み入れたい。 (data
augmentationでは一般化されているかどうか不明なため) ②パラメーター共有を普通のCNNに比べてさらに進めることで、 overfittingリスクを減らせる。
4.
Motivation <CNN・・・平行移動に対して強い> • Conv層・・・平行移動に対してequivariant • Pooling層・・・平行移動に対してinvariant → ①rotationに対してもinvariantを保ちたい。ネットワー クにそのinvariance/equivariantな性質を入れたい。 ②パラメーター共有を普通のCNNに比べてさらに進める ことで、overfittingリスクを減らせる。
5.
limitation ★Rotation:今回は90度の整数倍の回転に絞ってい る。 (30度などの鋭角回転を入れると、画像的な補間 が必要となり、さらに計算量が増えるから、とい う理由で除外している) ★鏡面に関して対象な画像(dihedral symmetry)に関 しても、実装は容易に可能だ、としている。
6.
基本的な考え方 画像を4方向に回転させたものを作成、同じ数のフィルタ を使いfeature mapを4倍作成できる
7.
提案手法 • 以下の3種類の層を組み込むことを提案 cyclic slicing,
cyclic pooling, cyclic rolling (cyclic stackingは実際には使っていない)
8.
①cyclic slicing, cyclic
pooling http://benanne.github.io/2015/03/17/plankton.html
9.
http://benanne.github.io/2015/03/17/plankton.html
10.
②cyclic rolling T(x )
11.
実験 <多クラス分類> • プランクトンデータセット (121クラス、95*95pixel、3037valid/27299train) • 銀河写真データセット (121クラス、95*95pixel、6157valid/55421train) <領域抽出> •
マサチューセッツの航空写真 (80*80pixel,137train/4valid/10test) Baseline CNNをコントロールとし、cyclic slicingなどを挿入すること でどれだけ性能が向上したかを示した。
12.
• Adam使用、プランクトンのみweight decayも併 用 •
Data augmentationは、ベースラインCNNにも、 筆者らの提案した手法にも、どちらにも使用し た
14.
実験その①(cyclic slice/pool)
15.
実験その②(cyclic rolling)
16.
<rollingの挿入の仕方> • Roll all(convの後にすべて挿入) •
Roll dense(dense layerの後にのみ挿入) • その後に1/2,1/4とついているのは、フィルタの 数をその倍率に縮小した、という意味。
18.
参考 • http://benanne.github.io/2015/03/17/plankton.ht ml • http://icml.cc/2016/reviews/871.txt
Download