Artificial Intelligence Terminology

Part 1 of a 4 part series on Artificial Intelligence

In this series of articles, we will provide a snapshot of the world of Artificial Intelligence (AI), from its terminology and origins to future trend predictions to help with understanding how this modern technology came to dominate the world’s conversation about our collective future.

Machine learning, deep learning, neural networks, supervised networks and more. These are some of the terms bandied about in articles and podcasts about AI.? We thought we’d start here – what do these terms mean in the context of a discussion about AI?

Let’s begin by understanding what Artificial Intelligence is at its core. IBM says, “Artificial intelligence, is technology that enables computers and machines to simulate human intelligence and problem-solving capabilities.”? When used independently or in conjunction with other technologies like sensors, geolocation, or robotics, AI can undertake tasks that would typically necessitate human intelligence or involvement. Now, let’s take it a step further and define some of the basic terms and concepts.

Machine Learning

The folks at Google Cloud like to say Machine Learning automatically enables a computer or system to learn and improve from experience rather than via explicit programming. Machine learning uses algorithms to analyze large amounts of data, acquire insights, learn from them and make informed decisions.? Machine learning is not about teaching a computer to mimic human intelligence, it is specifically aimed at teaching a computer to perform specific tasks on its own.

Deep Learning

Deep Learning is a subset of machine learning. Unlike machine learning, deep learning removes the pre-processing of data and relies on patterns found within. This allows the ingesting of big unstructured data such as an image or a sentence without human intervention, allowing a computer to perform more complex tasks such as driving vehicles and picking the fastest route on the fly as traffic conditions change.?

Neural Networks

Neural Networks are another important subset of machine learning and the backbone to deep learning. Inspired by how the human brain learns; Amazon likes to describe Neural Networks using interconnected nodes in a layered structure that signal each other as they pass data. This creates a system that is adaptive, creating an ability to learn from mistakes.? Thus, these artificial neural networks can attempt to solve complicated problems like summarizing documents or recognizing faces.?

Natural Language Processing

Natural Language Processing (NLP) assists computer systems with understanding and interpreting human language. Using deep learning algorithms, Natural Language Processing is used in tons of daily activities – chatbots, speech recognition, and machine translation.

Robotics

Robotics systems are a form of artificial intelligence that can replicate or substitute human action of controlling physical objects. Applications for this subset can be found in the manufacturing industry as they automate the manufacturing processes.

Expert Systems

Expert system (knowledge-based systems) is a subset of artificial intelligence that stimulates the decision-making capabilities of a human expert within specific domains. This is done by analyzing a knowledge repository made by domain experts, with formats varying from databases to semantic network. An inference engine is built which deduces information from the rules and data provided through the knowledge repository, providing recommendations like an expert in the field.

Computer Vision

Computer Vision is a subset of artificial intelligence that teaches computer systems to process information from visual inputs such as digital images or videos. The computer systems are then able to make predictions and identify defects or issues.

In part 2 of our series, we’ll provide a brief history and evolution of Artificial Intelligence.