Deep Learning Solutions about Artificial intelligence (AI)

Artificial Intelligence (AI) Solution- Although many definitions of artificial intelligence (AI) have emerged in recent decades, in a 2004 article (PDF, 106 KB) (link outside IBM)John Mc Cart hyprovides the following definition:“It is science and engineering that create systems intelligent machines,in particular intelligent computer programs. It’s about a similar task to using computers to understand human intelligence, but AI doesn’t have to be limited to biologically observable methods.

Artificial Intelligence (AI) Solution ~ Artificial Intelligence (AI) Solution Iturhsweb ~ Artificial Intelligence (AI) Solution in Digital Marketing

But decades before that definition, the birth of the artificial intelligence conversation was defined by Alan Turing’s seminal work Computing Machinery and Intelligence (PDF, 89.8 KB) (link external to IBM) published in 1950 by Turing in this article., often referred to as the “father of arithmetic,” asks the question, “Can machines think?”

From there he offers a test now known as the “Turing test”in which the questioner would try to distinguish between a computer answer and a text answer. Although this test has undergone many revisions since its publication, it remains an important part of the history of artificial intelligence as well as an ongoing concept in philosophy because it uses ideas from linguistics.

Stuart Russell and Peter Nor vi g later published Artificial Intelligence:

A Modern Approach (link outside IBM)and became one of the leading textbooks on the study of artificial intelligence. In it, they examine four possible goals or definitions of AI that distinguish between rational and thought-based computer systems and actions:

Human Approach:

systems that think like humans
systems that behave like humans

Perfect Approach:

Systems that think rationally
Systems that work rationally

Alan Turing’s definition would fall into the category of “systems that behave like humans”.”

In its simplest form, AI is a field that combines computation and robust datasets to enable  also includes the sub-areas of machine learning and deep learning, which are often mentioned in connection with artificial intelligence. These disciplines include artificial intelligence algorithms that aim to create expert systems that make predictions or classifications based on input data.

Artificial Intelligence (AI) Solution ~ Artificial Intelligence (AI) Solution Iturhsweb ~ Artificial Intelligence (AI) Solution in Digital Marketing

There is currently a lot of hype surrounding the development of artificial intelligence that is expected of any new technology that comes to market. As noted in the Gartner Hype Cycle (link external to IBM), product innovations such as self-driving cars and personal assistants”follow a typical innovation progression from over-enthusiasm through disillusionment to an understanding of the ultimate importance and role of innovation in a market or domain.

“As Lex Fridman notes here (01:08:05) (link external to IBM) in his 2019 MIT talk, we are at the height of inflated expectations, near the bottom of disappointment.

Types of AI: weak AI vs. strong AI

Weak AI,also known as Narrow AI or Narrow Artificial Intelligence (ANI),is trained by AI and focuses on performing specific tasks. Bad AI drives most AI around us.“Tight” might be a more accurate term for this type of AI since it’snot weak.

Apps like Apple’s Siri, Amazon’s Alexa, IBM’s Watson, and self-driving vehicles.

Strong AI consists of Artificial General Intelligence (AGI) and Super Artificial Intelligence (ASI). Artificial General Intelligence(AGI) or artificial general intelligence is a theoretical form of artificial intelligence in which a machine would be as intelligent as humans;It is said to have a confident awareness that has the ability to problem solve,learn and plan for the future.

Artificial Super Intelligence (ASI),also known as super intelligence,aims to surpass the intelligence and capabilities of the human brain. While the Force AI is still purely theoretical and there are currently no practical examples, that doesn’t mean that AI researchers aren’t exploring its development as well. Meanwhile, the best examples of UPSs might come from science fiction, like HAL, the superhuman rogue computer wizard in 2001: A Space Odyssey.

Artificial Intelligence Applications

Today there are many real-world applications of artificial intelligence systems.

Here are some of the more common examples:

Speech Recognition:

Also known as Automatic Speech Recognition (ASR), Computer Speech Recognition, or Speech to Text, and is a feature that uses Natural Language Processing (NLP) to convert human speech into a written format. Many mobile devices include voice recognition in their systems,for example to perform voice searches Siri–or make text messages more accessible.

Customer Service:

Virtual online agents replace human agents in the customer journey. They answer frequently asked questions (FAQs) on to pics such as shipping or offer personalized advice, cross-sell products or sizing suggestions for users and change the way we think about customer engagement on websites and social media platforms. Examples include messaging bots on e commerce sites with virtual agents, messaging apps like Slack and Facebook Messenger, and tasks typically performed by virtual and voice assistants.

Computer Vision: 

This artificial intelligence technology allows computers and systems to extract meaningful information from digital images, videos, and other visual data and to respond to those inputs.This ability to make recommendations sets them apart from image recognition tasks. Powered by convolutional neural networks, computer vision has applications in photo tagging in social media, radio logical imaging in healthcare, and self-driving cars in the automotive industry.

Recommendation Engines:

By mining data on past consumer behavior, AI algorithms can help uncover trends in the data that can be used to develop more effective cross-selling strategies. It is used to provide customers with relevant supplementary recommendations during the checkout process for online merchants.

Automated Stock Trading:

AI-powered high-frequency trading platforms designed to optimize stock portfolios execute thousands or even millions of trades daily without human intervention.

The History of Artificial Intelligence: Key Data and Names

The idea of ​​a“thinking machine” dates back to ancient Greece. But since the advent of electronic computing (and in relation to some of the topics covered in this article),significant events and milestones in AI development include:

1950: Alan Turing publishes Computing Machinery and Intelligence. In the article,Turing-known for cracking the Nazi ENIGMA code during WWII-proposes to answer the question “Do machines think?”

And introduced the Turing test to determine whether a computer could have the same intelligence (or intelligence scores) as a human. The value of the Turing test has been debated ever since.

Artificial Intelligence (AI) Solution ~ Artificial Intelligence (AI) Solution Iturhsweb ~ Artificial Intelligence (AI) Solution in Digital Marketing

1956: John McCarthy coined the term “Artificial Intelligence” at the first AI conference at Dartmouth College. (McCarthy invented Lisp). In the same year, Allen Newell, J.C. Shaw and Herbert Simon develop Logic Theorist, the first working artificial intelligence program.

1967: Frank Rosenblatt invents the Mark 1 Perceptron, the first computer based on a neural network that “learns” through trial and error. Just a year later, Marvin Minsky and Seymour Papert published a book entitled Perceptrons, which became both a seminal work on neural networks and, at least for a time, an argument against future research projects on neural networks of neurons.

1980: Neural networks using a back propagation algorithm for learning are widely used in AI applications.

1997: IBM Deep Blue then defeats world chess champion Garry Kasparov in a chess match (and a rematch).

2011: IBM Watson defeats champions Ken Jennings and Brad Rutter in Jeopardy!

2015: Baidu’s Minwa supercomputer uses a special type of deep neural network called a convolutional neural network to identify and classify images more accurately than the average human.

2016: Deep Mind’s Alpha Go program, powered by a deep neural network, defeats World Go Champion Lee Sodol in a five-game match. The victory is remarkable considering the large number of possible moves during the game (more than 14.5 trillion in just four moves!). Google later bought Deep Mind for $400 million.