Skip to main content

Posts

Showing posts from March, 2018

Reinforcement Learning - Introduction

This section of the course I'm taking was an introduction to reinforcement learning. Reinforcement Learning allows machines and software agents to automatically determine the ideal behaviour within a specific context (e.g. playing a game) in order to maximize performance. In a given environment, if the computer makes a good move, it receives a reward (e.g. a good score). If it makes a bad move, it receives a bad score. My code for this section can be found here . My overview of reinforcement learning can be found here . Below are the notes I took while watching the lectures for this section, then some observations I made while implementing the several programs. Notes What is Reinforcement Learning? Reinforcement Learning allows machines and software agents to automatically determine the ideal behaviour within a specific context (e.g. playing a game) in order to maximize performance Simple reward feedback is required for the agent to learn its behaviour; this is k...

Autoencoders - Introduction

This section of the course I'm taking covered autoencoders. An autoencoder is a simple neural network with the same number of inputs as outputs whose primary purpose is dimensionality reduction. For example, suppose we have a large set of data has 30 features (dimensions) and we want to reduce the number of features to have only 2 features. The autoencoder network determines which combinations of the original features are the most important, and constructs 2 new  features to represent them. The network then attempts to reconstruct the original data based on the newly encoded features. This is be extremely useful for image and video compression, but also for summarizing a set of data and determining the important features. I find that this property can best be summed up by the quote below from this Quora post by the user Luis Argerich: " Imagine our data has a vector for each user and each dimension corresponds to a movie the user likes or not (1 or 0 for each movie), th...

Deep Learning - Tensorboard and TensorFlow Abstraction APIs: Estimator, Keras, and Layers

In this section of the course I'm taking, I'll be learning about three of the most used TensorFlow Abstraction APIs: Estimator, Keras, and Layers. Abstractions allow for simplification of code by rolling many of the lines of code that we commonly write into a built-in function. My code for this section can be found here . While learning the Estimator API, I used a dataset of different types of wine. Each object had 13 attributes to describe it, most of them chemical. Each wine also had a label that could be either red, white, or rosé. I used the estimator.DNNClassifier API with three hidden layers of 13 neurons each. After the model was trained, it achieved a prediction accuracy (i.e. predicting if the wine is red/white/rosé given its attributes) of 97%. While learning the Keras API, I used the same dataset and data organization code as before. It took a couple lines fewer to solve the same classification task, but the average was slightly lower -- 91% compared to th...