Decoding the ResNet architecture

Thu, 02 Nov 2017 23:27:27 -0400

Introduction

Fast.ai’s 2017 batch kicked off on 30th Oct and Jeremy Howard introduced us participants to the ResNet model in the first lecture itself. I had used this model earlier in the passing but got curious to dig into its architecture this time. (In fact in one of my earlier client projects I had used Faster RCNN, which uses a ResNet variant under the hood.)

ResNet was unleashed in 2015 by Kaiming He. et.al. through their paper Deep Residual Learning for Image Recognition and bagged all the ImageNet challenges including classification, detection, and localization.

Network In Network architecture: The beginning of Inception

Fri, 20 Oct 2017 23:27:27 -0400

Introduction

In this post, I explain the Network In Network paper by Min Lin, Qiang Chen, Shuicheng Yan (2013). This paper was quite influential in that it had a new take on convolutional filter design, which inspired the Inception line of deep architectures from Google.

Motivation

Anyone getting introduced to convolutional networks first come across this familiar arrangement of neurons designed by Yann LeCun decades ago:

Fig. LeNet-5

Yann LeCun’s work (1989, 1998) triggered the convolutional approach, which takes into account the inherent structure of the incoming data (mostly image data) while propagating them through the network and learning about them.

Deep Learning on Teleported.in

Decoding the ResNet architecture

Introduction

Network In Network architecture: The beginning of Inception

Introduction

Motivation