Artificial Intelligence Explained

By Anthony Coggine January 31, 2018

Artificial intelligence is advancing rather rapidly. Smart software is ubiquitous, from programs that automatically sort through e-mail to computerized assistants that help us organize our schedules. Self-driving cars are beginning to populate our streets and anxiety over computerization is becoming a real threat to jobs. Things are moving more quickly and new applications of AI boggle the mind.

It seems like artificial powers a whole lot, but what is it exactly?

Artificial Intelligence Definition

In the simplest terms, artificial intelligence is intelligence originating from a machine. Strictly speaking, machines with AI should be aware of their environment and manipulate that environment to achieve their goals, just as humans do. AI is used much more loosely than that strict definition allows, however. Any machine that copies the logic and problem solving ability of the human mind is dubbed AI.

So, for example, while a personal assistant like Siri may not be aware of its environment, it effectively parses speech, extrapolates meaning, and mimics speech patterns. Siri’s mimicry is so effective and her capabilities so eclipse normal software programs that we informally call the program AI.

Researchers, computer scientists, and AI developers want to push AI much further than mimicry. There’s much more to AI than speech recognition. Self-driving cars are just one exciting possibility for smart software. These bigger and better applications of AI require more sophisticated solutions than Siri, however. This is where the amazing areas of AI research come into play: machine learning, deep learning, and general artificial intelligence.

Machine Learning

Artificial intelligence can only move forward if machines acquire the ability to learn things on their own. Machine learning is the process by which software is endowed with the ability to think for itself. Machines can self-improve, analyze datasets to come up with their own solutions, and apply algorithms to new sets of data without machine learning developers needing to type in new code.

Machine learning is more sophisticated than many applications we may consider artificial intelligence. Software that can learn does not need to rely on the preprogrammed algorithms of its creators. It can make its own conclusions. Machines learn from experiences as humans do, which allows for the analysis of massive datasets in record time. Already machine learning has been used to further cancer research, for example, through the use of big data and machine learning.

Deep Learning

Machine learning takes artificial intelligence a step further, and deep learning takes an extra leap.

You can think of deep learning is a subsection of machine learning. Deep learning attempts to recreate the human mind. Using mathematics, deep learning creates something called the neural network. Basically, this neural network attempts to recreate the interconnectedness of biological brain cells, or neurons.

A neural network learns through experience and stores knowledge through neural connections. Artificial neural networks replicate a network of brain cells through advanced mathematics.

Artificial General Intelligence

Artificial general intelligence, also known as “strong AI” or “full AI” is the ability for a machine to execute any task a human could. At present, artificial intelligence is able to perform one task or a group of related tasks very well, but AI does not have the ability to apply things they’ve learned to unrelated tasks. To put it another way, AI lacks common sense.

A human brain intrinsically has the ability to learn many different subjects, tasks, and skills. An underlying algorithm that replicates this process has yet to be created. If a general learning algorithm could be formulated, then general artificial intelligence could be possible.

Conclusion

Artificial intelligence is as simple concept: endow a machine with the mind of a human. Getting a machine to achieve general artificial intelligence is much more difficult than it may seem, however.

All the same, artificial intelligence continues to advance. In the decades to come, if machine learning and deep learning continue to push the field further, artificial intelligence could begin to rival the intellectual capabilities of a human being.

About the Author: Anthony Coggine is a HR professional turned business writer. He has been covering a range of topics including training, HR, recruiting and cryptocurrency news. 




Edited by Mandi Nowitz

Contributing Writer

SHARE THIS ARTICLE
Related Articles

Hughes: WAN Optimization Expertise, Homegrown Solution Differentiate SD-WAN

By: Paula Bernier    2/16/2018

The SD-WAN marketplace is a crowded one. But Hughes Network Systems says it brings unique expertise and proven technology to the table. And that, Jeff…

Read More

Juniper Security Expert: Behavior Analytics Helps Address Threat Complexity

By: Paula Bernier    2/16/2018

Organizations are changing their cybersecurity strategies, says Juniper Networks Cybersecurity Strategist Nick Bilogorskiy, who presented the closing …

Read More

Welbitz Wins ITEXPO's Idea SHOWCASE

By: Paula Bernier    2/16/2018

It was a sweep. Both the audience and the judges at ITEXPO's IDEA Showcase Thursday picked Welbitz as the winner. The company went up against fellow s…

Read More

Discussing Evolving Security Threats at ITEXPO

By: Gerald Baldino    2/15/2018

This afternoon at ITEXPO, HD Voice News Editor-in-Chief Doug Mahoney led a panel titled "How to Beat Evolving Security Threats," where he was joined b…

Read More

ITEXPO Keynote: Dialpad CEO Says Architecture Matters

By: Maurice Nagle    2/14/2018

DialPad CEO Craig Walker opines about the future of business communications, looking back to his first version of the company and to where the industr…

Read More