优先
- Few-shot Autoregressive Density Estimation: Towards Learning to Learn Distributions
- MAML
Summary
- MAML: meta-learn an initial condition
- LSTM optimization: meta-learn a good initial condition + LSTM-based optimizer
list
[x] Matching networks for one shot learning 3.6-3.7
[ ] Optimization as a model for few-shot learning. LSTM(一半)
[ ] Learning to Compare: Relation Network for Few Shot Learning
[ ] label efficient learning of transferable representations across domains and tasks
[ ] Meta-learning with memory-augmented neural networks (hard)
[ ] One-Shot Generalization in Deep Generative Models
[ ] meta learning with temporal convolutions(必看)
[ ] Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks(MAML)
[ ] Learning to learn by gradient descent by gradient descent
discrimtive methods
[34] Wang, Yu-Xiong, and Martial Hebert. Learning to learn: Model regression networks for easy small sample learning. European Conference on Computer Vision. Springer International Publishing, 2016.
[33] Bertinetto, L., Henriques, J. F., Valmadre, J., Torr, P., & Vedaldi, A. (2016). Learning feed-forward one-shot learners. In Advances in Neural Information Processing Systems (pp. 523-531).
[ ] Prototypical Networks for Few-shot Learning
[35] Triantafillou, Eleni, Hugo Larochelle, Jake Snell, Josh Tenenbaum, Kevin Jordan Swersky, Mengye Ren, Richard Zemel, and Sachin Ravi. Meta-Learning for Semi-Supervised Few-Shot Classification. ICLR 2018.(extend prototypical network)
[ ] 密度估计 Few-shot Autoregressive Density Estimation: Towards Learning to Learn Distributions
[ ] multimodal deep learning
[ ] cosegmentation
others
https://zhuanlan.zhihu.com/p/28639662
https://zhuanlan.zhihu.com/p/32270990
深度增强学习框架(trail and error)
终生学习Life Long Learning
少样本学习Few Shot Learning
多任务学习Multi Task Learning
多智能体学习Multi Agent Learning
学会学习Meta Learning/Learning to Learn
迁移学习Transfer Learning