AI is a long game. What is there right now is good enough, but in a couple of years, it will be very good. In about five years, it will be excellent. And in ten years, it will be impossible to work without it.
There’s no turn back, Artificial Intelligence (AI) will be part of our future. From resolving simple day-to-day pain points, such as house choirs or laundry and clothes folding, to more existential problems such as the cleaning of our oceans and the preservation of our environments, to the exploration of new worlds, it will shape the future of our society.
But what is AI and what does it means and how does it fits with the other buzzwords? This article will review the most commons subfields related to this topic to attempt to clarify how all work together.
Artificial Intelligence (AI)
Artificial Intelligence is the intelligence displayed by machines, in contrast with the natural intelligence displayed by humans and other animals, which perceives its environment and makes decisions to maximize chances of achieving its goal.
Machine Learning (ML)
Machine Learning is a subset of Artificial Intelligence that uses complex statistical techniques to enable machines to improve on tasks through experience. It’s the science to get computers to learn to do something, without being specifically programmed how to perform a task. The ML tasks can be using supervised learning, from labelled training data, or unsupervised learning, using unlabelled training data, which doesn’t produce evaluation of accuracy.
Sample training data can be downloaded as datasets from the UCI Machine Learning Repository.
Deep Learning (DL)
Deep Learning is part of a broader family of machine learning methods based on learning data representations, as opposed to task-specific algorithms.
(Artificial) Neural Network (ANN)
Artificial neural network (ANN) is an interconnected group of nodes (i.e. computer systems), inspired by the vast network of neurons that constitute animal brains, with each node representing an artificial neuron connected to other artificial neurons. It learns and improves progressively without prior knowledge by evolving a set of characteristics from learning materials (datasets).
Generative Adversarial Network (GAN)
Generative Adversarial Networks are a class of artificial intelligence algorithms used in unsupervised machine learning, implemented by a system of two neural networks contesting against each other via an adversarial process. Two models are trained simultaneously, a generative model G that captures the data distribution, and a discriminating model D that estimates the probability that a sample came from the training data. It is a powerful approach that has shown its strong ability creating and improving algorithm models generation after generation.
Natural Language Processing (NLP)
Natural Language Processing is concerned with the interactions between computers and human (natural) languages, and in particular with programming computers to develop an understanding of the large natural language data. Its a method of communicating using a natural language, such as English, with an intelligent system, such as a robot, to perform as instructions from a dialogue.
Data Science is the method to extract knowledge and insights using computational and statistical skills to solve data-driven problems. Rapidly expanding, it includes machine learning, deep learning, large-scale data analysis. It has applications in e-commerce, search/information retrieval, natural language modelling, finance, bioinformatics and related areas in artificial intelligence.
Robotics is an interdisciplinary branch combining computer science AI subfields, electrical engineering and mechanical engineering, to produce machines that operate autonomously to substitute or replicate humans. They are composed of sensors to receive feedback from the environment, actuators to interact, and AI to process information, control and take a decision.
Science-fiction author and professor of biochemistry Isaac Asimov is often given credit for being the first person to use the term robotics.
Machine Learning is a field that has grown out of AI, or Artificial Intelligence.
A machine learning is described by the following. A computer program is said to learn from experience E, with respect to some task T, and some performance measure P, if its performance on T as measured by P improves with experience E:
E = the experience
T = the task
P = the performance, i.e. the probability of correct answer
In general, any machine learning problem can be assigned to one of two broad classifications: Supervised learning and Unsupervised learning.
- Supervised learning : Learn by examples as to what a face is in terms of structure, color, etc so that after several iterations it learns to define a face.
- So, if you are training your machine learning task for every input with corresponding target, it is called supervised learning, which will be able to provide target for any new input after sufficient training. Your learning algorithm seeks a function from inputs to the respective targets. If the targets are expressed in some classes, it is called classification problem. Alternatively, if the target space is continuous, it is called regression problem.
- Unsupervised learning : since there is no desired output in this case that is provided therefore categorization is done so that the algorithm differentiates correctly between the face of a horse, cat or human (clustering of data)
- Contrary, if you are training your machine learning task only with a set of inputs, it is called unsupervised learning, which will be able to find the structure or relationships between different inputs. Most important unsupervised learning is clustering, which will create different cluster of inputs and will be able to put any new input in appropriate cluster.
The most common tasks of ML are regression and classification:
- Regression: the output variable takes continuous values. It involves estimating or predicting a response. Also, given a set of data, find the best relationship that represents the set of data.
- Classification: the output variable takes class labels.. It is identifying a group membership. Also, given a known relationship, identify the class that the data belongs to.
We can see that regression and classification start from opposing ends: to find a pattern or to find the pattern that it belongs to.