IS 130 - CH 14

Data Shift

A bias that comes from a mismatch between the data used to train and test the system and the data the system actually encounters in the real world.

Recurrent Neural Network

A neural network designed to access previous data such as sequential data or time-series data during iterations of input. Used in applications where its decision must be based on previous output such as moving a robotic arm, reading a sentence, predicting time series, and composing music.

Convolutional Neural Network

A neural network designed to separate areas of image inputs by extracting features to identify edges, curves, and color density and then recombine these inputs for classifications and predictions. It is used for facial recognition.

Algorithm

A problem-solving method expressed as a finite sequence of steps.

Intelligent Agent

A software program that assists users, or acts on their behalf, in performing computer-related tasks. A software more advanced than chatbots. They do not simply provide answers from a knowledge base. Rather, they must be able to do something for a user, such as understand customer intent.

Cluster Analysis

A technique to group, or segment, data points to identify common characteristics.

Information Agent

A type of intelligent agent that searches for information and display it to users

Reinforcement Learning

A type of machine learning in which the system learns to achieve a goal in an uncertain, potentially complex environment. The system faces a game-like situation where it employs trial and error to find a solution to a problem.

Strong AI

Also known as artificial general intelligence-is hypothetical artificial intelligence that matches or exceeds human intelligence. In other words, it refers to the intelligence of a machine that could successfully perform any intellectual task that a human being can.

Weak AI or Narrow AI

Artificial intelligence that performs a useful and specific function that once required human intelligence to perform, and does so at human levels or better. Like character recognition, speech recognition, machine vision, robotics, data mining, medical informatics, and automated investing.

Machine Learning (ML)

Is an application of artificial intelligence that provides systems with the ability to automatically learn and improve from experience without being explicitly programmed.

Generative Adversarial Network (GANs)

It consists of two neural networks that compete with each other in a zero-sum game in an effort to segregate real data from synthetic data. They separate real data from noise, it is also used for deep-space photography, inpainting, and in other applications where completing missing digital data is required.

False Positive

It is a result that indicates that a given condition exists when it in fact does not. Like convicting an innocent person, identifying an e-mail as spam when it is not, flagging a legitimate transaction as fraudulent, and many others.

Artificial Neural Network

It is a set of virtual neurons, or nodes, that work in parallel to simulate the way the human brain works, although in a greatly simplified form.

Deep Learning

It is a subset of machine learning in which artificial neural networks learn from large amounts of data.

Unsupervised Learning

It is a type of machine earing that searches for previously undetected patterns in a data set with no pre-existing labels and with minimal human supervision.

Semi-Supervised learning

It is a type of machine learning that combines a small amount of labeled data with a large amount of unlabeled data during training.

Classification

It refers to a predictive modeling problem in which the system generates a class label for a given set of input data. There are four types of classification: Binary classification, Multi-class classification, Multi-label classification, and Imbalanced classification.

Computer Vision

It refers to the ability of information systems to identify objects, scenes, and activities in images. Included in medical imaging to improve predicting, diagnosing, and treating diseases; facial recognition; and autonomous vehicles.

Natural Language Processing

It refers to the ability of information systems to work with text the way that humans do. For example, these systems can extract the meaning from text, and they can generate text that is readable, stylistically natural, and grammatically correct. It is used for analyzing customer feedback.

Types of Machine Learning

Supervised Semi-supervised Unsupervised Reinforcement

Speech Recognition

Technology that focuses on automatically and accurately transcribe human speech. This technology must manage diverse accents, dialects, and background noise.

Back Propagation

The process where the values of the weights of each pathway and the bias values of each node are slightly changed in anticipation that the next iteration of data flowing through the neural network will result in a smaller error, or loss, upon output.

Artificial Intelligence

The theory and development of information systems that are capable of performing a task that normally requires human intelligence.

Imbalanced Classification

These are classification problems in which the number of classes in each class is unequally distributed.

Binary Classification

These are classification problems that have only two class labels, like e-mail spam detection (spam or not).

Multi-Label Classification

These are classification problems that have two or more class labels, where one or more class labels can be predicted from each example.

Multi-class Classification

These are classification problems with more than two class labels, like news article categories, plant species classification, etc.

Expert Systems (ESs)

These are computer systems that attempt to mimic human experts by applying expertise in a specific domain.

Supervised Machine Learning

Where developers train the system with labeled input data and the expected output results. After the system is trained, developers feed it with unlabeled input data and examine the accuracy of the output data.

Monitoring and Surveillance Agents

also called predictive agents, constantly observe and report on some item of interest.

Chatbots

computer programs that you interact with via a chat interface. The program communicates with a user when a predetermined action occurs, such as a user typing a pot-up box or speaking to a device. Though, they do not learn from the interaction.

Signs of Intelligence

learning or understanding from experience, making sense of ambiguous or contradictory messages and responding quickly and successfully to new situations.