AI MOCK Ultimate Study Guide

Lonesome together

A person's partial, unknown-to-us or explicit lack of connection; Thoreau observed a city is often a place where many people are "lonesome together."

Covenantal divine fidelity

Hebrews 13:5: "Never will I leave you; never will I forsake you."

Immanuel

Matthew 1:22-23 meaning "God with us".

Faith

Confidence in what we hope for and assurance about what we do not see (Hebrews 11:1).

Zakar (remember)

Hebrew root meaning purposeful, interventionist action taken on someone's behalf at the appointed time.

Behavioral transmission

Proverbs 13:20 principle that walking with the wise makes one wise, while a companion of fools suffers harm.

First-Order Predicate Logic (FOPL)

Predicate logic expansion of propositional logic that decomposes propositions into objects and predicates and supports quantification.

Constant

A named entity in the domain, such as John or Pneumonia.

Variable

Placeholder symbol representing an arbitrary, unspecified domain entity, such as x or y.

Predicate

Relational symbol mapping entities to truth values, indicating a property or association, e.g. Student(x).

Function

Mapping operator that takes an input object and returns another specific object in the domain, e.g. MotherOf(x).

Universal quantifier

Symbol (∀) meaning an expression holds for all entities in the universe of discourse.

Existential quantifier

Symbol (∃) meaning an expression holds for at least one entity in the universe of discourse.

Conjunctive Normal Form (CNF)

A conjunction of clauses, each a disjunction of literals, used as the uniform representation for resolution.

Literal

An atomic predicate formula or its direct negation.

Clause

A disjunction (OR) of literals.

Unification

Process of finding variable substitutions (a unifier) that make two first-order expressions syntactically identical.

Resolution refutation

Inference method on CNF clauses that negates the goal, adds it to the knowledge base, and derives the empty clause to prove the goal.

Skolemization

Replacing existential variables with Skolem constants or Skolem functions to eliminate existential quantifiers.

Standardization apart

Renaming variables so that no variable name is shared across distinct formulas or quantifiers.

Horn clause

A clause containing at most one positive literal.

Selective Negative Linear Resolution (SNLR)

A goal-directed, backward-chaining linear resolution strategy for Horn clause knowledge bases, mirroring Prolog execution.

Fuzzy set

A set in which each element has a membership degree in [0, 1] defined by a membership function rather than binary membership.

Fuzzy union (OR)

Fuzzy union operator, defined as the maximum of the two membership grades.

Fuzzy intersection (AND)

Fuzzy intersection operator, defined as the minimum of the two membership grades.

Fuzzy complement (NOT)

Fuzzy complement operator, defined as 1 minus the membership grade.

Extension principle

Method for extending crisp functions to fuzzy inputs, using the supremum of membership over the pre-image of each output.

Fuzzification

Four-stage fuzzy controller step converting crisp inputs into membership grades.

De-fuzzification

Stage that converts the aggregated fuzzy output into a crisp control signal.

Centroid (Center of Gravity) method

Defuzzification method computing the balance point of the area under the aggregated membership curve.

Mean of Maximum (MOM)

Defuzzification method averaging all output values where the membership function reaches its maximum.

Mamdani inference

Mamdani-style fuzzy inference where the consequent is a fuzzy set, valued for interpretable linguistic outputs.

Sugeno (Takagi-Sugeno) inference

Fuzzy inference where the consequent is an analytical (often linear) function, efficient for adaptive control.

Statistical pattern recognition

Pattern recognition paradigm that classifies inputs using statistical models and probability theory from labeled data.

Syntactic (structural) pattern recognition

Pattern recognition paradigm that represents patterns using formal grammars, structural primitives, and relations.

Regression

Supervised learning task with continuous outputs, versus classification with discrete outputs.

k-Means clustering

Unsupervised learning method that partitions data into k clusters by iteratively assigning points to the nearest centroid and updating centroids as means.

Linear discriminant function

Linear function g(x) = w^T x + w_0 whose zero set forms the decision boundary separating two classes.

k-Nearest Neighbors (k-NN)

Instance-based classifier that assigns the majority label among the k closest labeled training points.

Nearest Neighbor Classifier (NNC)

Stored-sample supervised method that assigns the label of the single closest training point (k = 1).

Overfitting

Model that performs well on training data but poorly on test data because it fits training noise.

Underfitting

Model that lacks capacity to capture underlying patterns, performing poorly on both training and test data.

Activation function

Artificial neuron output computed as f(z), where z is the weighted sum of inputs plus bias.

Perceptron learning rule

Learning rule where weight change equals learning rate times target minus output times input, updating only on errors.

Delta rule

Gradient-descent learning rule for continuous outputs, forming the basis of backpropagation (Widrow-Hoff rule).

Backpropagation

Algorithm that propagates loss gradients backward through layers using the chain rule to update weights.

Learning rate

Hyperparameter governing the size of weight updates; too large causes divergence, too small causes slow convergence.