Types of Artificial Intelligence

Deep Learning
Deep Learning describes an algorithm that independently learns from inputs from its environment.

Sweetspot
While Deep Learning (DL) is the hottest topic in AI/ML, the use cases are limited due to the inability of Deep Learning systems to "understand the world." Instead, deep learning is an exercise in number crunching that can detect anomalies and trends in vast bodies of data.

Pros:
Autonomous learning

The more data is available, the more useful the results will be

Cons:
Needs vast bodies of data to work

Merely treats the environment as nondescript data patterns

Results are opaque to the human brain

Very narrow application

Results are not directly actionable but need a human brain or another cognitive technology for interpretation

Frequently Asked Questions

How much data do I need to effectively use Deep Learning?

How do I know upfront that my Deep Learning model works and how reliable it will be?

How broadly will the results be usable?

How do I judge reliability and accuracy of my Deep Learning model?

If Deep Learning does not understand the world, how could Deep Mind win at jeopardy?

What is Single Vector Deconstruction? No worries, there is a non-technical answer.