Artificial Intelligence Flashcards

Artificial Intelligence

• Concerns the creation of machines that exhibit intelligence.
• Encompasses various subfields including machine learning, robotics, and natural language processing.

Rational Agent

• Defined as an agent maximizing expected utility based on knowledge.
• Acts to achieve the best outcome in scenarios with uncertainty.

Turing Test

• Provides a framework for defining machine intelligence.
• Assesses whether a machine's behavior is indistinguishable from that of a human.

Natural Language Processing (NLP)

• Combines computer science and linguistics to enable machines to understand and interact using human languages.

Intelligent Agent

• Autonomous entities that perceive their environment and act towards achieving specified goals.

Knowledge Representation (KR)

• Translates information into symbols for effective inference and information creation.

Automated Reasoning

• Aims to understand various dimensions of thought processes through computer science and mathematical logic.

Machine Learning

• Involves algorithm design for computers to adapt behaviors based on empirical data from sensors and databases.

Computer Vision

• Involves methods for processing and understanding images and high-dimensional data to derive actionable information.

Robotics

• Focuses on designing and operating autonomous machines with sensory feedback and information processing capabilities.

Cognitive Science

• Interdisciplinary field integrating AI models and psychological techniques to study human cognition.

Syllogisms

• Logical structures providing patterns for valid arguments and conclusions based on premises.

Logic

• Explores valid reasoning strategies and identifies forms of logical fallacies.

Logicism

• Philosophical stance suggesting that mathematics is reducible to logic.

Bounded Rationality

• Concept stating that individuals' rational decision-making is limited by available information and cognitive capacity.

Rationalism

• Philosophy emphasizing reasoning as a key understanding tool, associated with thinkers like Descartes and Aristotle.

Dualism

• Descartes' belief in a non-physical component of the mind that exists outside of nature's laws.

Materialism

• Contrasts dualism, positing that the mind arises entirely from physical brain operations.

Empiricism

• Philosophy asserting knowledge originates from sensory experiences.

Induction

• Principle stating general rules emerge from repeated associations.

Logical Positivism

• Merges empiricism and rationalism to form a unified epistemological framework.

Confirmation Theory

• Explores how knowledge is gained through experiential evidence.

Algorithm

• Step-by-step procedures for performing calculations and solving problems.

Incompleteness Theorem

• Gödel's assertion on the limitations of formal axiomatic systems in arithmetic.

Computability

• Relates to problem-solving efficiency and correspondence to algorithms.

Intractability

• Describes theoretically solvable problems that are practically too complex to address.

NP-Complete (NP-C)

• Class of decision problems where solutions can be verified quickly but locating those solutions is not efficient.

Probability

• A fundamental mathematical concept crucial for AI, highlighted by Gerolamo Cardano's work.

Utility

• Mathematical representation of preferred outcomes, refined through the work of economists.

Decision Theory

• Integrates probability and utility theories to inform decision-making under uncertainty.

Game Theory

• Studies strategic decision-making, focusing on the interactions between rational agents.

Operations Research

• Applies analytical methods to improve decision-making and management practices.

Markov Decision Processes (MDPs)

• Mathematical framework for modeling decisions with both random and controlled outcomes.

Satisficing

• Strategy aiming for adequate solutions rather than optimizing for the best outcome.

Neuroscience

• Study of the brain and nervous system's structure and functions.

Neuron

• Specialized cell responsible for processing and transmitting information through electrical signals.

Technological Singularity

• Theoretical future point when machines surpass human intelligence, creating unpredictable advancements.

Behaviorism

• Approach dismissing mental processes based on the reliability of introspection.

Cognitive Psychology

• Explores how humans process internal mental activities, including perception and problem-solving.

Control Theory

• Engineering discipline focused on managing dynamical systems to maintain desired outputs.

Homeostatic Devices

• Mechanisms conceptualized to achieve stable adaptive behavior through feedback loops.

Cybernetics

• Research field established by Wiener, exploring control and communication in machines.

Objective Function

• Central goal in control theory aimed at maximizing a defined quantity over time.

Computational Linguistics

• Combines linguistics and computer science to model natural language computationally.

Hebb’s Principle

• Demonstrates how neurons adapt by modifying connection strengths during learning.

Physical Symbol System

• Composes symbols that convey information through structured manipulation.

Lisp

• Long-standing AI programming language characterized by its distinct syntax.

Microworlds

• Limited problem domains selected for AI development, focusing on specific intelligence tasks.

Adaline

• A basic neural network model consisting of weights and activation functions.

Perceptron

• Binary classification algorithm utilizing linear functions for decision-making.

Back-Propagation

• Common neural network training approach for refining outputs to minimize errors.

Genetic Algorithms

• Heuristic search methodologies inspired by natural evolutionary processes for optimization.

Neural Networks

• AI models mimicking brain functionality by interconnecting simpler units to achieve complex behaviors.

Expert Systems

• AI systems replicating human expertise to solve complicated problems through reasoning.

Certainty Factors

• Uncertainty calculus introduced in MYCIN for medical diagnosis assessments.

Frames

• AI data structures representing stereotyped situations to form comprehensive knowledge concepts.

Connectionist Models

• Approaches modeling cognition and behavior as emergent from networks of interconnected units.

Hidden Markov Models

• Statistical models assuming unobserved states in a Markov process framework.

Data Mining

• Discovery of patterns in large data sets using AI, machine learning, and statistical methods.

Bayesian Network

• Probability model representing variable dependencies through a directed acyclic graph.

Human-Level AI

• Concept aiming to develop machines capable of human-like learning, thinking, and creativity.

Strong AI

• Represents AI with capabilities matching or exceeding human cognitive skills.

Nouvelle AI

• 1980s research approach focusing on insect-level intelligence viability over human-level performance.

Situated Approach

• Agent design methodology emphasizing environmental interaction over abstract reasoning.

Embodied Cognition

• Cognitive processes simulating bodily functions directly rather than through logical abstractions.

AI Winter

• Periods of reduced interest and funding in AI research, characterized by cycles of hype and disappointment.### Inference
• Inference involves drawing logical conclusions from premises that are known or assumed to be true.
• The conclusion reached through inference is referred to as an idiomatic.

Deductive Reasoning

• Deductive reasoning is the process of moving from general statements to reach a logically certain conclusion.
• It operates by using true premises to arrive at a conclusion that is also true.

Inductive Reasoning

• Inductive reasoning constructs or evaluates propositions based on observations from individual instances within a specific class.
• This reasoning is distinct from arriving at a general conclusion through specific examples.

Analogy

• Analogy is a cognitive process that transfers information or meaning from one subject (the source) to another (the target).
• It highlights connections between different concepts by drawing parallels.