# Miscellaneous

Collection of code, talks, links, and other interesting things

## Projects & Code

During my life as a student, computer engineer and scientist, I have been involved in many different projects and written thousands lines of code which may interest other people. Then, I decided to share some of this content. Find below project descriptions and source code for some of my projects:

## CUDA Prog. Exercises

C++ self-study parallel-computing

Collection of programming exercises of the book Programming Massively Parallel Processors A Hands-on Approach by David Kirk and Wen-mei W. Hwu. In this self-study project, we implement 2D convolution, matrix multiplication, and other simple mathematical operations using C++ and CUDA framework.

## Image Processing Algs.

java coursework undergrad

This is a collection of traditional digital image processing algorithms, e.g., image filtering, restoration and morphological operations, implemented in JAVA. This is not an optimized code, since this is a self-study project developed for educational purpose.

## ML Notebooks

python ipython-notebooks self-study

A collection of traditional machine learning algorithms and applications ranging from density estimation to classification and regression techniques. These algorithms are coded in python using IPython Notebooks which allows to join text explanation, mathematical formulation and code. This is a self-study project with educational purpose.

## CVX Notebooks

python ipython-notebooks self-study

A collection of convex optimization problems and applications ranging from portfolio optimization to doubly stochastic matrix approximation. These algorithms are coded in python using CVX and IPython Notebooks which allows to join text explanation, mathematical formulation and code. This is a self-study project with educational purpose.

## CNN Object Classifier

python talks RVSS 2017

This project was developed for the CNN workshop of the Robotic Vision Summer School 2017. This workshop taught how to train a CNN classifier for user-defined classes using keras and tensorflow. Using this code, you can try different CNN's architectures and preprocessing pipelines in your own data.

## Deep Activity Classifier

python experiments ReScience

We experiment three different deep learning models to classify activities in videos. We start with the independent frame prediction baseline, then we add temporal information using RNNs or 3D convolutions. These experiments are coded in python using Keras and Tensorflow frameworks.

## Useful links

- Coding: Numpy , Keras , Tensorflow .
- Reading: Nurturing Science , Karpathy's Blog .
- Tools: Timezone Converter , AI Conference deadlines .