Elements of AI

अब Quizwiz के साथ अपने होमवर्क और परीक्षाओं को एस करें!

Layer

All the neurons here get inputs from neurons on the previous one and feed their output to the next one.

John Searle

American philosopher best known for his work in the philosophy of language—especially speech act theory—and the philosophy of mind. Formerly a professor emeritus in UC Berkeley's Department of Philosophy, had his emeritus status revoked as of June 19, 2019, along with all the privileges of that title, following a determination that he violated university policies against sexual harassment and retaliation.

Cycle of hype

An AI boom in optimism followed by and an AI winter.

Alan Turing

An English mathematician and logician, considered to be the father of computer science.

Thomas Bayes

An English statistician, philosopher and Presbyterian minister who is known for formulating a specific case of the theorem that bears his name. Never published what would become his most famous accomplishment; his notes were edited and published after his death by Richard Price.

Synapses

Each output signal wire may be connected to one or more input signal wires at these intersections.

Robotics

Building and programming vehicles that have at least some level of autonomy and include sensors and actuators so that they can operate in complex, real-world scenarios.

Probability

Can be quantified (expressed as a number) and it can be right or wrong from an AI perspective.

Odds

Can be thought of as a single fraction or a ratio.

Neural networks

Can mean either a "real" biological system such as the one in your brain, or an artificial system simulated in a computer.

Convolutional neural networks (CNNs)

Can recognize the object anywhere in the image no matter where it has been observed in the training images using their key property that they can detect image features such as bright or dark (or specific color) spots, edges in various orientations, patterns, and so on.

Deep learning

Certain kinds of machine learning techniques where several "layers" of simple processing units are connected in a network so that the input to the system is passed through each one of them in turn. This architecture has been inspired by the processing of visual information in the brain coming through the eyes and captured by the retina. Allows the network to learn more complex structures without requiring unrealistically large amounts of data.

John McCarthy

Coined the term Artificial Intelligence in 1956 and is often referred to as the Father of AI.

Reinforcement learning

Commonly used in situations where an AI agent must operate in an environment and where feedback about good or bad choices is available with some delay.

Euler Diagram

Consists of shapes that corresponds to concepts, which are organized so that overlap between the shapes corresponds to overlap between the concepts.

Bayes classifier

Machine learning technique that can be used to assign objects into two or more classes. Trained by analyzing a set of training data, for which the correct classes are given.

Unsupervised learning

No labels or correct outputs used and the task is to discover the structure of the data.

AI boom

Paradigm of high optimism each time an all-encompassing, general solution to AI has been said to be within reach.

Common saying in statistics

"all models are wrong, but some are useful" aphorism generally attributed to statistician George E.P. Box

Value for lower bound for all probability estimates to fix those with zero counts

1/100,000

1:13 Convert from odds to probability expressed as a fraction.

1/14

6:5 Convert from odds to probability expressed as a percentage.

54.5%

6:5 Convert from odds to probability expressed as a fraction.

6/11

1:13 Convert from odds to probability expressed as a percentage.

7.1%

George E.P. Box

A British statistician, who worked in the areas of quality control, time-series analysis, design of experiments, and Bayesian inference. He has been called "one of the great statistical minds of the 20th century".

General AI

A machine that can handle any intellectual task

European General Data Protection Regulation

A major step towards transparency. It requires that all companies that either reside within the European Union or that have European customers must: - Upon request, reveal what data they have collected about any individual (right of access) - Delete any such data that is not required to keep with other obligations when requested to do so (right to be forgotten) - Provide an explanation of the data processing carried out on the customer's data (right to explanation)

Weights

A set of adaptive parameters used as multipliers on the inputs.

Face2Face

A system capable of identifying the facial expressions of a person and putting them on another person's face in a Youtube video.

MNIST dataset

A team led by Yann LeCun releases a mix of handwritten digits from American Census Bureau employees and American high school students. Became a benchmark for evaluating handwriting recognition.

Turing Machine

A theoretical model that can compute anything that is computable and led to the invention of programmable computers.

Lyrebird

A tool for automatic imitation of a person's voice from a few minutes of sample recording. While the generated audio still has a notable robotic tone, it makes a pretty good impression.

Narrow AI

AI that handles one task.

Robot

Any kind of vehicles that have at least some level of autonomy and include sensors and actuators.

Chinese Room Thought Experiment

Argument that even if a machine behaves in an intelligent manner, it doesn't follow that it is intelligent or that it has a "mind" in the way that a human has.

Posterior odds

Assessment of the odds after obtaining the information.

Prior odds

Assessment of the odds before obtaining some new information that may be relevant.

AI methods

Automated reasoning based on the combination of perfectly understandable principles and plenty of input data, both of which are provided by humans or systems deployed by humans.

De-anonymization

Breaking the anonymity of data that we may have thought to be safe.

Yann LeCun

French computer scientist working primarily in the fields of machine learning, computer vision, mobile robotics, and computational neuroscience. Received the 2018 Turing Award. Sometimes referred to as one the "Godfathers of AI" and "Godfathers of Deep Learning"

Minimax Algorithm

Given a state of the game, simply computes the values of the children of the given state and chooses the one that has the maximum value if it is Max's turn, and the one that has the minimum value if it is Min's turn.

Supervised learning

Given an input and and the task is to predict the correct output or label.

Base rate fallacy

If presented with related base rate information (i.e., general information on prevalence) and specific information (i.e., information pertaining only to a specific case), people tend to ignore the base rate in favor of the individuating information, rather than correctly integrating the two.

Turing Test

Imitation game that if a computer is indistinguishable from a human in a general natural language conversation, then it must have reached human-level intelligence.

Optimal Moves

In a game tree, at any Min node (where it is Min's turn), the optimal choice is given by the child node whose value is minimal, and conversely, at any Max node (where it is Max's turn), the optimal choice is given by the child node whose value is maximal.

"Child" Node

In a game tree, on the successive levels, the possible states that can result from the previous player's moves.

"Root" Node

In a game tree, the beginning position, usually depicted at the top of the diagram.

Paradigm

In philosophy of science, the term used for a trend.

AI winter

Interest in the field faltered and research efforts were directed elsewhere as the true nature of the remaining problems dawned after years of struggling and unsatisfied promises and pessimism about the paradigm accumulates.

Weak AI

Philosophical distinction of a machine acting intelligently; systems that exhibit intelligent behaviors despite being "mere" computers.

Strong AI

Philosophical distinction of a machine being intelligent; possessing a "mind" that is genuinely intelligent and self-conscious.

Transitions

Possible moves between one state and another.

Likelihood ratio

Probability of the observation in case the event of interest divided by the probability of the observation in case of no event.

Parallel processing

Process many pieces of information at the same time. Graphics processors (or graphics processing units, GPUs) have this capability and they have become a cost-effective solution for running massive methods simultaneously.

Linear regression

Produces a numerical prediction that is not constrained to be an integer. Adds up the effects of each of the feature variables to produce the predicted value. Better suited in situations where the output variable can be any number.

Data Science

Recent umbrella term (term that covers several subdisciplines) that includes machine learning and statistics, certain aspects of computer science including algorithms, data storage, and web application development.

Costs

Refer to the fact that, oftentimes the different transitions aren't all alike. They can differ in ways that make some transitions more preferable or less desirable.

Perceptron

Simple neuron model with the step activation function. Called the "mother of all artificial neural networks" because was among the very first formal models of neural computation.

Neurons

Simple units that receive and transmit signals to each other. Most of the time, they do nothing but sit still and watch for signals coming in.

Convolutional layer

Special kind that is a very elegant solution to the problem of too many weights.

Value alignment problem

Specifying the objectives of the system so that they are aligned with our values is very hard.

Deep Learning

Subfield of machine learning where the complexity of a mathematical model and that the increased computing power of modern computers has allowed researchers to increase this complexity to reach levels that appear not only quantitatively but also qualitatively different from before.

Machine Learning

Systems that improve their performance in a given task with more and more experience or data.

Logistic regression

Take the output from another type of regression, which is a number, and predict label A if the output is greater than zero, and another label B if the output is less than or equal to zero. Can also give us a measure of uncertainty of the prediction.

Suitcase Word

Terms that carry a whole bunch of different meanings that come along even if we intend only one of them coined by Marvin Minsky, a cognitive scientist and one of the greatest pioneers in AI.

Adaptivity

The ability to improve performance by learning from experience.

Autonomy

The ability to perform tasks in complex environments without constant guidance by a user.

Foxes

The category of people who have many small ideas as classified by the political scientist Philip E. Tetlock, author of Superforecasting: The Art and Science of Prediction. Tend to be clearly better at prediction, especially when it comes to long-term forecasting.

Hedgehogs

The category of people who have one big idea as classified by the political scientist Philip E. Tetlock, author of Superforecasting: The Art and Science of Prediction.

Algorithmic bias

The embedding of a tendency to discriminate according ethnicity, gender, or other factors when making decisions about job applications, bank loans, and so on. Isn't a hypothetical threat conceived by academic researchers. It's a real phenomenon that is already affecting people today.

Game Tree

The nodes are arranged in levels that correspond to each player's turns so that the starting node (usually depicted at the top of the diagram) is the beginning position.

State Space

The set of possible situations.

Axons

The wire that transmits the outgoing signal.

Dendrites

The wires that provide the input.

Superintelligent AI

Unrealistic doomsday scenario where AI can outsmart humans emerges as an unintended result of developing AI methods. Will not emerge from developing narrow AI methods and applying them to solve real-world problems.

Singularity

Unrealistic doomsday scenario where a system that optimizes and "rewires" itself so that it can improve its own intelligence at an ever accelerating, exponential rate. Such superintelligence would leave humankind so far behind that we become like ants that can be exterminated without hesitation.

Backpropagation

Uses automatic differentiation method developed in the Master's thesis written by Seppo Linnainmaa and was later applied by other researchers to quantify the sensitivity of the output of a multilayer neural network with respect to the individual weights.

Collaborative filtering

Uses other users' data to predict your preferences and you only see content recommended for you.

Generative adversarial networks (GANs)

Using two networks compete against each other. One of the networks is trained to generate images like the ones in the training data. The other network's task is to separate images generated by the first network from real images from the training data.

Nearest neighbor classifier

When given an item, it finds the training data item that is most similar to the new item, and outputs its label.

Bayes rule

posterior odds = likelihood ratio × prior odds


संबंधित स्टडी सेट्स

ACSM CPT Chapter 12 Client Fitness Assesments

View Set

Anatomy 2 - Digestive system Exam

View Set

Vocab Level G Unit 13-15 Mastery

View Set

Chapter 39: Nursing Care of the Child With an Alteration in Sensory Perception/Disorder of the Eyes or Ears

View Set