Deep Learning Paper Suggestions

[sumodm]

We will add the initial recommendations here and then move them to a papers.md as we start organizing the papers.

Application and Specific Architectures

1. Hypercolumns for Object Segmentation and Fine-Grained Localization
2. Deep SimNets

Theory

3. On the Expressive Power of Deep Learning: A Tensor Analysis

Recurrent Papers

4. LSTM: A Search Space Odyssey

[prabhuteja12]

Just adding a list of papers that we can discuss and the finalized ones can be added to the homepage.

Week-3: Maybe this page for the basics. And AlexNet

Week-4: VGG Net and Going Deeper with Convolutions

Week-5: This year's ILSVRC winner http://arxiv.org/pdf/1512.03385.pdf

[utkarshsimha]

Batch Normalization Deep Residual Networks

[sumodm]

Saliency

• NUS: Deep Learning for Saliency
• Saliency Detection by Multi-Context Deep Learning
• Large-Scale Optimization of Hierarchical Features for Saliency Prediction in Natural Images

Visualization

• Deep Inside Convolutional Networks: Visualizing Image Classification Models and Saliency Maps

Deeper Architectures

• Deep Residual Learning for Image Recognition

Shallow Architectures

• Highway Networks

Misc

• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift

Context

• A Critical View of Context: Not Deep Learning

[sumitbinnani]

Limits of Computer Image Recognition

Atoms of Recognition in Human and Computer Vision http://www.pnas.org/content/early/2016/02/09/1513198113.full.pdf

Blog Post on the aforementioned paper http://arstechnica.com/science/2016/02/tiny-blurry-pictures-find-the-limits-of-computer-image-recognition/

[utkarshsimha]

He Initialization for Relu

[IMG-PRCSNG]

Fractal Nets - Alternative to Residual Nets