Machine Learning And Its Impacts

Machine Learning And Its Impacts

Over the past few decades, there have been many definitions of artificial intelligence (AI), but in 2004, John McCarthy proposed the following definition: Computers to grasp human intellect, but AI need not be restricted to biologically observable ways.

However, Alan Turing’s groundbreaking work Computing Machinery and Intelligence, which was published in 1950, marked the beginning of the artificial intelligence discussion decades before that formulation. Rice paddies the “Turing test,” in which human interviewers try to differentiate between text responses from computers and those from people.

Artificial Intelligence: A Modern Approach, which was later published by Stuart Russell and Peter Norvig, quickly rose to prominence as one of the most important textbooks in the field. They go into his four potential objectives or definitions of AI that set computer systems apart based on reason, thought, and deed.

The Human Approach:

 Human-like Computer Systems

A machine that acts like a person

The Best Course of Action:

System That Reasons Logic

A mechanism that operates logically.

The term “a human-like system” would be used to describe Alan Turing’s definition.

Artificial intelligence can be defined as a field that combines computation and large data collections to facilitate problem solving. It also encompasses the machine learning and deep learning subfields, which are frequently used interchangeably with artificial intelligence. These fields use artificial intelligence (AI) methods to build expert systems that classify or forecast based on incoming data.

As with any new technology entering the market, there is still a lot of hype surrounding the development of AI. Self-driving cars and personal assistants, for example, follow “a normal innovation path, from exuberant enthusiasm through a time of ineptitude to an awareness of the relevance and role of innovation in a market or industry,” according to Gartner’s hype cycle. We are at the height of heightened expectations and are moving toward the bottom of disillusionment, as Lex Fridman emphasized in his MIT 2019 address.

Types of artificial intelligence

         Weak AI Vs Strong AI

Artificial intelligence (AI) that is trained and aimed to carry out particular tasks is referred to as weak AI, narrow AI, or narrow artificial intelligence (ANI). The majority of the AI we encounter today is weak. This form of AI is everything but weak; it supports some incredibly powerful applications, including Apple’s Siri, Amazon’s Alexa, IBM Watson, and self-driving cars. “Narrow” could be a more realistic descriptor.

Artificial General Intelligence (AGI) and Artificial Super Intelligence are the two components of strong AI (ASI). A machine is considered to have intelligence comparable to that of a human being in general artificial intelligence (AGI), or AI in general. He is stated to have self-awareness, the capacity for problem-solving, learning, and future planning. Super intelligence, commonly referred to as artificial super intelligence (ASI), is supposed to be more intelligent and capable than the human brain. Although there are no current applications for strong AI and it is currently completely theoretical, this does not mean that researchers are not looking into its development. The best instances of ASI, like as HAL, can be found in science fiction.

Machine Learning And Deep Learning

Knowing the differences between deep learning and machine learning is advantageous because the terms are sometimes used interchangeably. In addition to being fields of artificial intelligence, deep learning is also a sub field of machine learning, as was already mentioned.

Neural networks are the actual building blocks of deep learning. A neural network with three or more layers, which includes inputs and outputs, is referred to as “deep” in deep learning. This neural network can be thought of as a deep learning algorithm. The diagram that follows provides an overall illustration of this.

Each algorithm learns differently in deep learning and machine learning. Deep learning greatly automates the feature extraction portion of the process, removing some of the need manual human involvement and enabling the usage of larger datasets. Deep learning can be conceived of as “scale able machine learning,” as Lex Fridman noted in the same MIT talk that was mentioned above. Traditional or “deep” machine learning is primarily reliant on human input. To identify differences between data sources, human specialists create feature hierarchies, but they typically require learning from more structured data.

Labeled datasets, also known as supervised learning, can be used in “deep” machine learning to guide algorithms, but they are not necessarily necessary. It can automatically establish hierarchies of characteristics that set apart various types of data from unstructured material in its raw form (text, photos, etc.). Machine learning can be expanded in more interesting ways than human learning because it doesn’t require human intervention to process the data.

Leave a Reply

Your email address will not be published. Required fields are marked *