-
## 一言でいうと
Graph Convolutionをプロダクションレベル(30億ノード!)で使用したという金字塔的な論文(Pinterestで使われている)。グラフサイズが膨大なので近傍ノードをサンプリング(ランダムウォーク)で収集しており、収集はCPU・畳み込みはGPUで役割分担しMapReduceで分散処理して高速化している。
### 論文リンク
https://arxi…
-
## 一言でいうと
CNNの構造を自動設計する研究。まずランダムに構造を作ってもCIFAR-10で誤差率6~7%は出てしまうことを確認(既存の自動設計手法と同等の結果)。次いで、ネットワークの変形手法を4つに定式化し、何個か変化させる⇒ベストを採用しそれをベースにまた何個か変化、と繰り返していくNASHという手法を提案
![image](https://user-images.gith…
-
Define You:
- [x] Hacktoberfest2022 Participant
- [ ] Contributor
A convolutional network for face detection. This work is based on "Multi-view Face Detection Using Deep Convolutional Neural…
-
Define You:
- [x] Hacktoberfest2022 Participant
- [ ] Contributor
A convolutional network for face detection. This work is based on "Multi-view Face Detection Using Deep Convolutional Neural…
-
What are the differences and connections between percentage of all bands and contamination rate. thank you
-
# NLP : Character-level Convolutional Networks for Text Classification, 2016 | PaperCat
[http://localhost:4000/2020-08/paper2](http://localhost:4000/2020-08/paper2)
-
What do you think of adding Graph Convolutional Networks in a new section on ML and deep learning?
Here are some references:
* https://tkipf.github.io/graph-convolutional-networks/
* https://gi…
-
A deep convolutional generative adversarial network implemented in PyTorch! The project is designed to generate realistic images from random noise using the power of deep learning.
This project illus…
-
- 목적: CNN backbone 구조를 효율적으로 Scale up하고 싶음.
- CNN 모델을 Scale up 하려는 시도는 많이 이루어졌음. 그러면 성능이 더 좋아져서.
- 일반적으로 3가지 차원을 건드려 모델을 키움
- **depth** : 이건 layer 수를 늘리는 것. receptive field가 커버하는 범위가 넓어지기 때문…
-
# MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications #
- Author: Andrew G. Howard, Menglong Zhu, Bo Chen, Dmitry Kalenichenko, Weijun Wang, Tobias Weyand, Marco Andr…