Projects

Neural Network Knowledge Distillation

TensorFlow demonstration of the Knowledge Distillation framework to show how soft labels act as regularizers and a neural network’s knowledge can be transfered to a simpler model

Disjunctive Normal Networks

PyTorch Implementation with sklearn-like API of a Neural Network that constructs a Boolean Function by dividing the feature space with convex polytopes.

ML From Scratch

Machine Learning implementations from scratch. Using minimal dependencies this collection intends to cover fundamental machine learning algorithms: from linear regression to neural networks

Underactuated Robotics

Python implementations of Optimal Control and Optimization algorithms for simulated underactuated systems and walking robots

Gaussian Processes and Bayesian Optimization

A Python implementation of the Gaussian Processes framework with Bayesian Optimization. Fit noiseless or noisy data and use existing or custom kernels. Bayesian optimization module using existing acquisition functions (μ+kσ,...

Planar robotic arm with arbitrary number of rotary joints

2D robotic arm with arbitrary number of links and rotational joints. Implementation of trajectory planning and kinematic control.

Overused phrases in Suits TV show

Scraping and processing of the TV series Suits transcripts to extract, analyze and visualize most common phrases

Learning from Demonstration with Dynamic Movement Primitives

Minimal Python implementation of the Dynamic Movement Primitives (DMPs) framework for the description of demonstrated trajectories with dynamical systems.