# 論文筆記 - Network Architecture

# Survey

# Neural Architecture Search: A Survey

## [论文笔记系列 - Neural Network Search ：A Survey](https://yq.aliyun.com/articles/661799)

---

# Google DeepMind, [Progressive Neural Networks](https://hackmd.io/s/SJoGTU8yN), 2016

---

# Tianqi Chen, Ian Goodfellow, Jonathon Shlens, [Net2Net: Accelerating Learning via Knowledge Transfer](https://hackmd.io/s/rkv4MrLkV), ICLR'16, 2015

### [GitXiv](http://www.gitxiv.com/posts/GgBNprwm8xZ3Tzfwh/net2net-accelerating-learning-via-knowledge-transfer)

- 讓 CNN 根據需要 自動**拓展**架構
- 不屬於 compression 範圍

---
# [Network morphism](https://hackmd.io/s/BJVjy5vk4), ICML 2016

---
# Modularized Morphing of Neural Networks, 2017
- 針對 Network morphism 的問題所做的研究?

---

# Learning without forgetting, ECCV 2016, TPAMI 2017

---
# Overcoming Catastrophic Forgetting by Incremental Moment Matching, NIPS 2017

---

# Lifelong Learning with Dynamically Expandable Networks, ICLR 2018

---

# Simple and efficient architecture search for convolutional neural network, 2017

[网络结构搜索（3） —— Simple and efficient architecture search for convolutional neural network](https://blog.csdn.net/Lucifer_zzq/article/details/83315430)

---
# Efficient Multi-objective Neural Architecture Search via Lamarckian Evolution
- 允许收缩体系结构的近似网络结构来减少架构大小

---

# Taskonomy: Disentangling Task Transfer Learning
- [CVPR 2018 Best Paper Taskonomy 作者解读](https://zhuanlan.zhihu.com/p/38425434)

###### tags: `paper notebook`