marcelomata/adventures-in-deep-learning

State-of-the-art Deep Learning publications, frameworks & resources

Adventures in deep learning

State-of-the-art Deep Learning publications, frameworks & resources

Overview

Deep convolutional neural networks have led to a series of breakthroughs for in large-scale image and video recognition. This repository aims at presenting an elaborate list of the latest state-of-the-art works on the field of Deep Learning since 2013.

This is going to be an evolving repository and I will keep updating it (at least once every two weeks).

---

State-of-the-art papers (Descending order based on Google Scholar Citations)

1. Very deep convolutional networks for large-scale image recognition (VGG-net) (2014) [pdf] [video]
2. Going deeper with convolutions (GoogLeNet) by Google (2015) [pdf] [video]
3. Visualizing and Understanding Convolutional Neural Networks (ZF Net) (2014) [pdf] [video]
4. Deepface: closing the gap to human-level performance in face verification (2014) [pdf] [video]
5. Deep learning (2015) [pdf]
6. Fully convolutional networks for semantic segmentation (2015) [pdf]
7. Batch normalization: Accelerating deep network training by reducing internal covariate shift (2015) [pdf]
8. Delving deep into rectifiers: Surpassing human-level performance on imagenet classification (PReLU) (2014) [pdf]
9. Deep residual learning for image recognition (ResNet) by Microsoft (2015) [pdf] [video]
10. Spatial pyramid pooling in deep convolutional networks for visual recognition (SPP Net) (2014) [pdf] [video]
11. Fast R-CNN (2015) [pdf]
12. Faster R-CNN: Towards real-time object detection with region proposal networks (2015) [pdf]
13. Generative Adversarial Nets (2014) [pdf]
14. Understanding deep image representations by inverting them (2015) [pdf]
15. Spatial Transformer Networks (2015) [pdf] [video]

Classic publications

• ImageNet Classification with Deep Convolutional Neural Networks (AlexNet) (2012) [pdf]
• Rectified linear units improve restricted boltzmann machines (ReLU) (2010) [pdf]

Theory

1. Deep Neural Networks are Easily Fooled: High Confidence Predictions for Unrecognizable Images (2015) [pdf]
2. Distilling the Knowledge in a Neural Network (2015) [pdf]
3. Deep learning in neural networks: An overview (2015) [pdf]

Books

• Deep Learning Textbook - An MIT Press book (2016) [html]
• Learning Deep Architectures for AI [pdf]
• Neural Nets and Deep Learning [html] [github]

Courses / Tutorials (Webpages unless other is stated)

• Caffe Tutorial (CVPR 2015)
• Tutorial on Deep Learning for Vision (CVPR 2014)
• Introduction to Deep Learning with Python - Theano Tutorials
• Deep Learning Tutorials with Theano/Python
• Deep Learning: Take machine learning to the next level (by udacity)
• DeepLearnToolbox – A Matlab toolbox for Deep Learning
• Stanford Matlab-based Deep Learning
• Stanford 231n Class: Convolutional Neural Networks for Visual Recognition [github]

Resources / Models (GitHub repositories unless other is stated)

• VGG-net
• GoogLeNet
• ResNet - MatConvNet implementation
• AlexNet
• Fully Convolutional Networks for Semantic Segmentation
• OverFeat
• SPP_net
• Fast R-CNN
• Faster R-CNN

Frameworks & Libraries

• Caffe: [homepage] [github] [Installation Instructions]
• Torch: [homepage] [github]
• Theano: [homepage] [github]
• Tensorflow: [homepage] [github]
• Keras: [homepage] [github]

License

MIT