Blog


Convolutional Neural Networks




Tutorial overview with detailed visualizations of the evolution of convolutional neural networks up to the state-of-art. Find it here

Index

1: LeNet - LeCun - 1998 Paper (TBI)

2: AlexNet - Krizhevsky - 2012 Paper

3: GoogLeNet (Inception) - Szegedy - 2014 Paper (TBI)

4: VGG - Simonyan / Zisserman - 2014 Paper (TBI)

5: ResNets - He - 2015 Paper
  - Tutorial on Residual Deep Convolutional Networks (Demistifying ResNet Paper)
  - Tutorial on modifying the ResNet for CIFAR-10
    - PyTorch Implementation Explanation
    - Code for ResNets on CIFAR10
6: DenseNets - Huang - 2016 - Paper
  - Tutorial on Residual Deep Convolutional Networks (Demistifying ResNet Paper)
  - Tutorial on modifying the DenseNet for CIFAR-10
    - Code for DenseNets on CIFAR10

7: MobileNets - Howard - 2016 Paper (TBI)

8: FractalNets - Larsson - 2016 Paper (TBI)

9: SE-Nets - Hu - 2017 Paper







In this project I built a Neural Style Transfer Model to replicate the style of an image and copy it into an image we would like to preserve the content. I have transfered Fornite videogame appearence to Retiro Park, "the Central Park of Madrid", my city, in Spain.

Check out the documentation
Or directly go to the detailed explained PyTorch implementation .
Keras implementation is to be implemented.
Find the code on my GitHub







In this project, I have created a robot that lands in the center of a grid world, and performs a Simulatenous Localization And Mapping.
This world will include the agent, and some trees that will act as the landmarks.

This agent will have the ability to measure within a range where the trees are with a gaussian noise induced in the measurements, aiming to replicate inexactitude in real worl sensors.
Furthermore, the robot has the ability to randomly move within the world and keep sensing. This will help the robot update its representation of its belief states to improve the accuracy of its perception. The movement is also affected by gaussian noise in the displacement, to replicate imperfections of the real world robots

Since this projects relies on Graph SLAM approach, the beliefs are registered and computed into the matrices Omega and Xi.
Read it here




Semantic Segmentation and Scene Understading




- Theoretical Introduction
- Practical Implementation (TBI)