-
Will you consider to add to this library some similarities among neural network representations? Like CCA or CKA for example.
-
1.HoloPose: Holistic 3D Human Reconstruction In-The-Wild(2019)
mixture-of-experts rotation prior, part-based modeling( features co-varies with joint position)
code: can not open [http://arielai.com/…
-
From [Algorithmic Simplicity](https://www.youtube.com/@algorithmicsimplicity):
- [x] [Why Does Diffusion Work Better than Auto-Regression? - YouTube](https://www.youtube.com/watch?v=zc5NTeJbk-k)
-…
-
### 論文へのリンク
[[arXiv:1711.00937] Neural Discrete Representation Learning](https://arxiv.org/abs/1711.00937)
### 著者・所属機関
Aaron van den Oord, Oriol Vinyals, Koray Kavukcuoglu
- DeepMind
##…
-
## Abstract
- propose Vector Quantised Variational AutoEncoder (VQ-VAE)
- generative model that learns discrete representations
- prior is learnt rather than static
- solves the issue of "po…
-
- https://arxiv.org/abs/1711.00937
- 2017
監督なしで有用な表現を学習することは、機械学習における重要な課題である。
本論文では、このような離散的な表現を学習する、シンプルで強力な生成モデルを提案する。
我々のモデル、VQ-VAE(Vector QuantisedVariational AutoEncoder)は、2つの重要な点でVAEとは異なる…
e4exp updated
3 years ago
-
## 一言でいうと
VAEを利用して離散の潜在表現を学習する試み。Encoderの出力に近いベクトルを埋め込み空間から検索し、そこからDecoderで復元する。分布はEncoderの出力に近いところに1、それ以外が0であるone-hotな分布として定義される。これにより離散表現を獲得できるだけでなく、Decoderが強力すぎる場合に潜在表現が学習されない問題を克服している。
![ima…
-
# Neural Discrete Representation Learning #
- Author: Aaron van den Oord, Oriol Vinyals, Koray Kavukcuoglu
- Origin: https://arxiv.org/abs/1711.00937
- Related:
- https://avdnoord.github.io/…
-
http://science.sciencemag.org/content/360/6394/1204.full
- Deepmind
DL輪読会
- https://www.slideshare.net/DeepLearningJP2016/dlgqn-111725780
TMats updated
6 years ago
-
## 一言でいうと
シーンを認識で、A地点/B地点からの画像から、C地点からの画像を推定するというタスクを解かせた研究。各地点からの画像をネットワークで潜在表現に圧縮し、クエリ(C地点)と潜在表現から画像を生成する。これにより2Dから3Dへのレンダリングや、強化学習での学習効率の向上が行えたとのこと。
### 論文リンク
https://deepmind.com/blog/neu…