Complete AI guide with use cases
Introduction
Popularity of Artificial Intelligence (AI) has exploded in the past few years mainly due to technologies such as Big Data, Cloud Computing, GPUs and advancement in AI algorithms.
Big Data technologies like Hadoop and Spark have made data collection, data analysis and data cleaning easier. Cloud Computing makes deploying models cheaper while GPUs shortens the time it takes to train a model. Advancements in machine learning especially in deep learning frameworks has increased the accuracy of the models.
Today, anyone, anywhere in the world with an Internet connection can leverage the concepts in this article to create their own AI. AI can be used to predict anything, the only condition being that you have the relevant data (and lots of it!)
Machine learning
Machine learning is a branch of statistics which analyzes data and generates patterns in it. There are three categories of machine learning; descriptive, predictive and prescriptive.
The three major types of machine learning are; supervised, unsupervised and reinforcement leaning.
Supervised learning
What it is
When to use it
How it works
Popular Algorithms
Linear Regression
Logistic Regression
Linear/quadratic discriminant analysis
Decision tree
Naive Bayes
Support vector machine
Random forest
AdaBoost
Gradient-boosting trees
Simple neural network
Unsupervised learning
What it is
When to use it
How it works
Popular Algorithms
K-means clustering
Gaussian mixture model
Hierarchical clustering
Recommender system
Reinforcement learning
What it is
When to use it
How it works
Business Use Cases
Deep Learning
Deep learning or deep neural networks is a branch of machine learning which use an interconnected layer of calculators known as neurons which ingest large amounts of data in its input layers, digests it in its hidden layers and produce results in its output layers. Deep learning techniques are widely used by most, if not all major IT companies as their accuracy especially in image, voice and facial recognition is leaps and bounds above traditional machine learning algorithms.
There are many types of deep learning frameworks; boltzmann machine, autoencoders, capsule networks, general adversarial networks etc. In this article, we will focus on convolutional and recurrent neural networks.