Ch.1.4 help desk the impact of artificial intelligence

Artificial intelligence is best described as

A branch of computer science that builds systems to do tasks we associate with humans

Perception:

AI systems have senses just as we do. Instead of just five senses, AI systems use sonar, accelerometers, infrared, magnetic, and many other electronic sensors to gather data. Being able to combine all the data pouring in from sensors and then construct information from it is a difficult challenge.

Planning:

AI systems need to set goals and then achieve them. An AI system might need to plan how to move a blue block out of the way to reach a red one or how to rotate as it flies to fit through a narrow opening.

The Turing test

Is used to decide whether a computer can be considered intelligent.

Law Enforcement

AI-powered software enables detectives to analyze a wealth of information for similarities between crimes to detect patterns that may reveal serial crimes. A law enforcement specialty called computer forensics is growing in importance to fight modern crime. Computer forensic technicians analyze computer systems with specific techniques to gather potential legal evidence.

Recommendation engines

Are used to predict future purchases (or selections) from past activity.

Education

Speeding up grading of written assignments are AI powered robo-readers. These software agents can interpret and effectively grade written essays. Modern plagiarism agents use machine learning to spot similar patterns in writing and estimate the likelihood that it was plagiarized.

The Turing Test

Alan Turing was an early pioneer in computer design and cryptography. He proposed a simple test to distinguish between a human and a computer system. The Turing test places a person in a room, asking written questions of two other "people." But one person is a live human being (human foil) while the other is a computer system. If the questioner can't tell which one of the respondents is a computer, then Turing felt the computer had reached "intelligence." The computer is permitted to be deceptive; however, the human foil is required to help the questioner reach the correct conclusion.

Problem solving:

Human beings tend to make intuitive judgments when solving a problem, rather than performing a step-by-step analysis. AI programming combines a rules-based approach along with trying to make judgments with incomplete information.

Knowledge representation:

Knowledge representation involves encoding information about the world into formats that the AI system can understand. Humans possess a vast collection of general knowledge based on their experiences in the world. AI systems need to build knowledge bases to solve problems. Developing a knowledge base and using it efficiently are active areas of research and have been demonstrated effectively by IBM Watson.

Learning:

Like humans, AI algorithms improve through experience. Supervised learning is used when the system can be trained with a huge number of examples. When you teach Siri to recognize your voice by reading to it, you're using supervised learning. Unsupervised learning is when a system can look at data on its own and build rules for deciding what it's seeing.

Natural Language Processing (NLP):

NLP works to develop software that understands written and spoken words and can interact with us using human language.

Areas of active research in AI include

NLP, knowledge representation, perception (data gathering)

Social media

Social media giants like Facebook are already using AI algorithms to do facial recognition on images you post. But now software can classify images into categories like "trees" or "cars" automatically. Text can be analyzed to determine the intent of the author, allowing specific types of posts to be automatically deleted.

Which of the following statements is true about unsupervised learning?

There is not a set of training data supplied. The system builds its own set of rules for its decision-making. The algorithms improve their accuracy over time.

Artificial intelligence

is a branch of computer science that focuses on creating computer systems or computer-controlled machines (including robots) that have an ability to perform tasks that are usually associated with human intelligence and are (or were) generally performed by humans. By this definition, any computer program (or computer-controlled device) that accomplishes something normally thought of as intelligence by humans would be considered AI.