Concepts starting with R

219 concepts, each with a plain-language explanation, examples and diagrams. Pick a letter or open a concept.

Rademacher Complexity — The chaining technique uses covering numbers to bound Rademacher complexity.
Rademacher Complexity and Generalization Bounds — Rademacher complexity is a measure of the complexity of a set of functions or vectors.
Rademacher Complexity for Linear Predictors — Hard-SVM's output w_S is defined by Equation (26.19) and satisfies L_S(w_S) = 0 under the theorem's assumptions.
Rademacher Complexity of Linear Classes — The H1 proof reduces a Rademacher-complexity calculation to a finite-set problem.
Raising and catching custom exceptions — The else block in exception handling runs only when the try block succeeds without raising an exception.
Raising and Handling Exceptions — Custom exceptions are user-defined classes that inherit from Exception, enabling you to represent errors specific to your…
Raising Exceptions — The try..except statement separates normal code (try block) from error-handling code (except blocks).
Random Color Generation with the HSB Model — A JPanel subclass such as RandomStringsPanel draws its content inside the paintComponent() method, which the source describes as…
Random Number Generation in Java — The main program fills the 16 by 20 grid with random colors first, then runs a while loop that repaints the current square and…
Random Projection — The lemma gives a probabilistic guarantee that distances between vectors can be preserved after projection into a…
Random Variables — D is the distribution; z ∼ D describes sampling from it.
Random-Sample Q-Planning — Incremental planning breaks planning into small steps instead of requiring one uninterrupted calculation.
random.choices() for Multiple Random Selections — random.choice() selects one element at random from a sequence by generating a random index and returning the element at that…
random.randint() for Random Integers — random.choice() selects one element at random from a sequence by generating a random index and returning the element at that…
Randomization in Algorithms — A shuffling algorithm aims to put the elements of an array into a random order, with the typical case being shuffling a deck of…
Randomness in Programs — SimpleGraphicsStarter.java is the text's first example of a GUI program, meaning a program whose result is a graphical interface…
Range and Null Space — Eigenvectors are non-zero vectors that a matrix transforms only by scaling.
Ranking — Bipartite ranking is a two-label ranking problem involving relevant and non-relevant elements.
Ranking Algorithms and Search Engine Fundamentals — PageRank is computed by running sprank.py, which iteratively calculates importance scores based on the link structure in your…
Ranking and Complex Prediction Problems — Multiclass classification is classification into one of several possible target classes.
Ranking Functions — A linear ranking function applies the same vector w to instances through f(x) = w · x.
Ranking Performance Measures — Multiclass prediction can be reduced to binary classification, but the reduction must include a rule for combining binary outputs.
Ranking Problems — A multiclass predictor must combine evidence for several possible classes into one output.
Rate Limiting and Throttling Strategies — API keys are the mechanism vendors use to identify users and monitor their service consumption.
Reading and Interpreting Error Messages — User-supplied comments in backup filenames provide context and improve file organization when managing multiple backups.
Reading and Processing Files — A histogram is a frequency count stored in a dictionary, where keys are unique values (like email addresses) and values are…
Reading and Understanding Error Messages — but that is much uglier and error-prone. Second, the conversion to string would be done automatically by the format method…
Reading and writing byte streams — A copy program must use the byte streams InputStream and OutputStream because it cannot assume the file holds human-readable data.
Reading and Writing Files — Invisible characters like spaces, tabs, and newlines can cause silent errors in file operations.
Reading and Writing Files in Binary Mode — Binary files must be accumulated in a bytes buffer and cannot be printed as they arrive like text files.
Reading and Writing Files with Subclasses of InputStream and OutputStream — Byte streams (InputStream and OutputStream) move binary data without translation, while character streams (Reader and Writer)…
Reading and Writing Text Files — The newline character (\n) is a single invisible character that represents the end of a line, despite appearing as two symbols…
Reading and Writing Text Files in Python — Main memory (CPU and RAM) is where programs execute and store data while running, but it is volatile—everything is erased when…
Reading Basic Data Types With Scanner Methods — The Scanner class, in package java.util, was introduced to make reading basic data types from a character input source easier in…
Reading Chapter-Based Syllabi — The supplied material describes a 14-week graduate introduction course.
Reading Characters from a Reader — BufferedReader is a Reader that buffers input from a character-input stream, and any Reader can be wrapped in one to read full…
Reading Data Back from a File — A version number in a file's first line tells a program which layout rules were used, and a higher number signals a newer format.
Reading Data from Files: The read() and readline() Methods — The write() method puts data into a file exactly as provided and returns the character count written.
Reading Element Attributes — The reader starts from a default background and an empty list of curves, then changes them as it walks the DOM, and only assigns…
Reading Element Content from a Document — A DOM representation of an XML document is created in Java with two statements: one builds a DocumentBuilder through…
Reading Entire Files into Memory — The file handle does not contain file data; it is a connection to the file on disk that the for loop uses to read data…
Reading Entity-Relationship Diagrams — Primary and foreign keys are omitted from simplified data model diagrams because their placement follows an absolute…
Reading Error Messages Effectively — Random walk programming—making random code changes without a hypothesis—is inefficient and often masks the real problem instead…
Reading File Contents — Attempting to open a non-existent file raises a FileNotFoundError and prevents the program from obtaining a file handle.
Reading File Contents with File Methods — Opening a file asks the operating system to locate the file by name and verify it exists and is readable.
Reading Files and Iterating Over Lines — Parsing is the process of finding interesting lines in a file and extracting specific pieces of information from them.
Reading Files in Python — A text file can be conceptualized as a sequence of lines, analogous to how a Python string is a sequence of characters.
Reading Files Line by Line — The continue statement skips the rest of the current loop iteration and jumps back to the loop start, enabling efficient…
Reading Files Line-by-Line with a for Loop — read() loads an entire file into a single string, including all newline characters, and exhausts the file handle resource after…
Reading Files with UTF-8 Encoding — Use io.open with mode='wt' and encoding='utf-8' to write Unicode text to files.
Reading JSON from Files with json.load() — json.loads() transforms a JSON string into a Python data structure by parsing the text and building corresponding lists…
Reading Measurements with Single Units — Combined measurements such as '3 feet 7 inches' require one subroutine to read several number and unit pairs from a single line…
Reading Numbers From the User With TextIO — TextIO reads from a file have no sensible recovery path, so when a read fails it throws an exception rather than fixing the…
Reading Objects with ObjectInputStream — ObjectOutputStream is optimized to avoid writing the same object more than once; on a second encounter it writes only a…
Reading Per-Class Evaluation Results — An eval set measures a model's performance, so its composition and labels shape the score you see.
Reading Python Official Documentation — Square brackets in documentation indicate optional arguments; they are not part of the actual code.
Reading Python Tracebacks — An AssertionError is raised when an assert statement's condition is False, and it appears in the traceback with the file name…
Reading Research Papers Effectively — arXiv is a platform where recent research papers are listed, and the cs.LG recent listing is a practical way to see new machine…
Reading Source References Critically — The source pack names the phrase Trading Margin for Moat and attributes it to an industry analyst of a16z, dated June 2025, with…
Reading Text Data with Readers in Java — Buffering improves I/O efficiency by moving data in large chunks, which suits some devices better than reading individual bytes…
Reading Text Files with urllib — Binary files require byte-for-byte copying from a remote URL to your local disk.
Reading Text from Input Streams — A URL object identifies a resource; calling openConnection() on it returns a URLConnection.
Reading Traces for AI Debugging — The per-trace loop starts by reading individual sampled traces with their actual input and actual output, rather than the…
Reading Training Data Tables Critically — A token is a unit of text that a model reads, and dataset sizes are commonly reported as token counts, which roughly track the…
Reading User Input — raw_input() captures text from the keyboard and returns it as a string to your program.
Real-Time Optimization — An adaptive controller changes refinery operating parameters in real time.
Real-Valued Prediction — Linear regression is used to model relationships between explanatory variables and a real-valued outcome.
Realizability Assumption — Realizability demands perfect agreement between some hypothesis in H and a target labeling function.
Realizable and Nonrealizable Learning — Finite classes contain a limited number of hypotheses, often because their representations are bounded.
Realizable Assumption — PAC learnability is a definition involving a hypothesis class H, a function m_H, and a learning algorithm.
Realizable Binary Classification — Realizable ERM for homogeneous halfspaces can be expressed as a feasibility problem.
Realizable Compression Schemes — Compression in the unrealizable setting seeks minimum error within H, not necessarily zero error.
Realizable Learning — Axis-aligned rectangles are rectangles in R^n whose sides are parallel to the axes.
Realizable PAC Learnability — An ϵ-net turns a sample into a coverage guarantee for all sufficiently probable sets in a hypothesis class.
Reasoning About Assignment Statements — A program is correct when it produces the correct result as specified, so correctness is always measured against a specification.
Receptive Fields in Convolutional Networks — Subsampling replaces local groups of feature-map values with averaged summaries.
Recognizing and Handling Infinite Loops — The break statement exits a loop immediately, skipping any remaining code in the loop body.
Recurrent Artificial Neural Networks — Feedforward describes the absence of loops in an ANN's connections.
Recurrent Networks with Keras — The supplied section places bidirectional recurrent layers within advanced techniques for recurrent neural networks.
Recurrent Neural Networks — A language model learns the statistical structure of language by modeling possible next tokens.
Recurrent Neural Networks for Sequence Modeling — Character-level generation learns the relationship between a fixed-length character context and the character that follows it.
Recursion — A separate thread is useful for animation when the thread runs a recursive algorithm and the display must be redrawn many times…
Recursion Basics: Base Case and Recursive Case — The Towers of Hanoi problem is easy to solve recursively because the problem contains smaller copies of itself: move N minus 1…
Recursion in Threads — A Finish button cannot directly stop a recursive algorithm, because the click can arrive many levels deep in the recursion and…
Recursion: Functions Calling Themselves — Nested function calls occur when one function invokes another during its execution, causing the first function to pause while…
Recursive Definitions in Practice — A recursive definition uses the thing being defined as part of its own definition, and the source says this can be a very…
Recursive Relationships of Value Functions — The action-value function qπ(s, a) evaluates a specific action in a specific state under a policy.
References and Variables in Java — An object is garbage when no variable refers to it, and the system reclaims its memory
References, Pointers, and Null in Java — Some languages do not require variables to be declared, so a misspelled name can silently create a new variable; Java requires…
Regression — Supervised learning connects inputs with known targets through a collection of examples.
Regression Testing for Prompt Changes — An eval set is a living artifact that must change as the team learns more about the customer's domain.
Regret Analysis — An online convex optimization problem is defined by a convex hypothesis class, a domain, and a loss function convex in its first…
Regret Bound — Online Gradient Descent applies a repeated prediction-and-update process to online convex learning problems.
Regret in Online Learning — A hypothesis class with only two constant predictors is still vulnerable to an adversary that reacts to the learner's current…
Regular Expression Basics: Character Classes and Anchors — grep stands for Generalized Regular Expression Parser. It is a command-line program built into Unix, Linux, and macOS systems…
Regular Expression Flags and Options — Regular expressions are search strings with special characters that communicate matching criteria and extraction rules to a…
Regular Expression Fundamentals — Parentheses in regex patterns create capturing groups that extract specific substrings from a match.
Regular Expression Quantifiers and Repetition — grep stands for Generalized Regular Expression Parser. It is a command-line program built into Unix, Linux, and macOS systems…
Regular Expression Syntax and Anchors — A grep simulator reads a file line by line, tests each line against a user-supplied regular expression, and counts how many…
Regular Expressions — Always use raw strings when dealing with regular expressions. Otherwise, a lot of backwhacking may be required. For…
Regular Expressions for Text Splitting — A tokenizer class implements encode and decode, and it holds a vocabulary that maps text tokens to integer IDs along with an…
Regularization — Convexity of a learning problem has two requirements.
Regularization and Overfitting — Tikhonov regularization augments the loss with λ ‖w‖2.
Regularization in Deep Learning — Hyperparameters are architecture-level choices, such as layer count, units or filters, activation functions, and dropout rates…
Regularization in Machine Learning — Soft-SVM preserves the margin constraint while allowing controlled violations.
Regularized Learning Rules — Stability is about keeping output changes small when input changes are small.
REINFORCE Algorithm — This practice section applies REINFORCE to gridworlds.
REINFORCE Algorithm and Properties — The convergence argument for REINFORCE is based on the expected update, not on every individual episode.
REINFORCE eligibility vector — A Bernoulli-logistic unit is a stochastic neuron-like unit for binary actions.
REINFORCE Policy Gradient Algorithm — Exploration is necessary because a team needs varied collective actions in order to learn.
REINFORCE: A Monte Carlo Policy Gradient Method — The convergence argument for REINFORCE is based on the expected update, not on every individual episode.
Reinforcement Learning Agents and Environments — Do not identify the agent by tracing the outside of a robot's or animal's body.
Reinforcement Learning and Accumulated Reward — Expected reward describes an average, but it does not by itself describe variance or risk.
Reinforcement Learning and Animal Decision Processes — Model-free and model-based learning differ according to whether a model of the environment is used in decision-making.
Reinforcement Learning and Animal Learning — The defining question is whether reinforcement depends on the animal's behavior.
Reinforcement Learning and Artificial Intelligence — Reinforcement learning is connected to both mathematical disciplines and the sciences of biological learning.
Reinforcement Learning and Behavioral Control — Habitual behavior is primarily triggered by antecedent stimuli.
Reinforcement Learning and Neuroscience — Reinforcement learning concerns learning through interactions between an agent and its environment.
Reinforcement Learning and Policy Improvement — Instrumental conditioning is learning in which consequences modify behavior.
Reinforcement Learning and Prediction — A TD error expresses a prediction mismatch in a form used to update predictions.
Reinforcement Learning and Reward Signals — Values forecast total reward over the future, while prediction errors describe mismatches between expectations and observations.
Reinforcement Learning and Synaptic Weights — Synaptic efficacy describes how strongly neurotransmitter released at a synapse influences the postsynaptic neuron.
Reinforcement Learning Applications — The section broadens reinforcement learning beyond its standard ideas by examining relationships with psychology and neuroscience.
Reinforcement learning episodes — Episodic Semi-gradient Sarsa learns a value function by updating value-function weights θ during episodes.
Reinforcement Learning for Acrobot Control — Sarsa controls the acrobot through repeated online interaction rather than through access to a simulator model.
Reinforcement Learning for Recommendation Policies — Greedy optimization asks which recommendation is most likely to produce the next click.
Reinforcement Learning Fundamentals — The survey extends reinforcement learning beyond its standard foundations.
Reinforcement Learning Methods — A policy gradient method searches over parameter-defined policies rather than treating a value estimate as the primary search…
Reinforcement Learning Optimality — Value functions evaluate expected returns under a specified policy.
Reinforcement Learning Problem Overview — The Getting Started module contains the Getting Started chapter.
Reinforcement Learning Problems — The central question is what model of the environment is available to the agent.
Reinforcement Learning Updates — Reinforcement learning systems need generalization for broad AI and engineering applicability.
Reinforcement Learning vs Supervised Learning — Unsupervised learning is concerned with finding hidden structure in unlabeled data.
Relational Operators in Java — Boolean operators form complicated conditions from boolean values.
Relaxing the Realizability Assumption — True error evaluates a prediction rule h with respect to labeled points drawn according to the probability distribution D.
Relevance-Based Search — Ranking problems arrange instances by relevance rather than assigning an isolated label to one instance.
ReLU Operations — An element-wise operation applies a rule independently to each tensor entry.
Remote-First Engagements — The Scoping Phase is the first phase of a project and lasts a few days.
Removing Elements by Value with remove — pop removes an element by index and returns it; del removes an element by index without returning it.
Removing Nodes from a Linked List — Traversal moves through a linked list by following next pointers from the head until null is reached.
Removing Punctuation from Text — The Python split() function preserves punctuation attached to words, treating 'soft!' and 'soft' as different tokens and…
Replacing the Output Layer for Classification — The output tensor from the model has three dimensions: batch, one row per input token, and embedding values per token.
Representation Learning — Deep learning extends machine learning by emphasizing successive learned representations.
Representation Learning with Multiple Layers — The hypothesis space is the set of possible transformations or solutions available to a learning algorithm.
Representational Power in Neural Networks — Recurrent dropout applies dropout in recurrent layers to fight overfitting.
Representing Policies for Continuous Actions — A normal density describes how probability is distributed across possible real-valued actions.
Requesting Exceptions — A framed exception request asks for a one-time, time-bounded exception, not a permanent change to the standard process.
Requirements for AI System Deployment — The first filter for an AI project is whether the problem is one of the customer's top five priorities; the source says that if…
Research Directions in Reinforcement Learning — Part III is the book's final move beyond standard reinforcement learning ideas.
Researching a Company Before an Interview — The recruiter screen is the first filter, so it deserves real preparation rather than a casual approach.
Reserved Keywords — Identifiers are names given to variables and other entities; they must follow strict syntax rules.
Reserved Words and Keywords in Python — Exit the Python interpreter by typing quit() at the >>> prompt—the parentheses are essential.
Residual Connections — A layer graph can branch and merge, but it must remain acyclic except for processing internal to recurrent layers.
Responding to Menu Selections — A JMenuBar is a container for menus with no name and a no-parameter constructor; menus are placed in it with add(JMenu).
Responding to Mouse Events in a GUI Program — A MouseEvent provides the cursor position through evt.getX() and evt.getY(), returned as integers measured from the top left…
Responsibility Division in Object-Oriented Design — MVC divides a graphical component into Model, View, and Controller so that each object has one clearly defined role.
Responsible Deployment of Reinforcement Learning — Optimization unpredictability is one part of the wider challenge of responsible reinforcement learning deployment.
RESTful APIs and Data Exchange — XML (eXtensible Markup Language) is a text-based format for encoding structured, hierarchical data using custom tags that…
Retrieving Data with SELECT — INSERT adds new rows to a table; specify the table, columns, and values.
Retrieving Results from Futures — A Runnable is handed to an ExecutorService with execute, which returns as soon as the task is placed in a queue.
Return Statements and Return Values — The return statement is used to return from a function i.e. break out of the function. We can optionally return a value from the…
Return Values — The return statement is used to return from a function i.e. break out of the function. We can optionally return a value from the…
Return Values and Function Output — Flow of execution is the order in which statements are executed, always proceeding from top to bottom.
Return Values and How Functions Communicate Results — A function call redirects execution from the caller to the function body, then returns to the caller after the function completes.
Return Values and Passing Data Between Functions — Nested function calls occur when one function invokes another during its execution, causing the first function to pause while…
Return Values and the Return Statement — The return statement exits a function and optionally sends a value back to the code that called it.
Returning Values from Functions — The return statement is used to return from a function i.e. break out of the function. We can optionally return a value from the…
Returning Values from Methods — A linear search checks each item of an array in turn. It finishes as soon as it finds the item, and it can conclude the item is…
Returns and Episodes in Reinforcement Learning — The methods differ only in how they select returns from episodes with repeated visits.
Returns and Expected Discounted Rewards — Monte Carlo state-value estimation learns from complete observed returns rather than from a single outcome.
Returns and Monte Carlo Prediction — TD(0) and Monte Carlo can disagree because they use different evidence from the same episodes.
Returns and State Values — Episode data records what happened after states were observed.
Returns and Value-Function Updates — λ identifies important endpoint cases of the λ-return.
Returns in Reinforcement Learning — Continuing tasks describe agent-environment interactions that do not naturally divide into identifiable episodes.
Reusable Parts and Clean Templates — A slice library is a moat for solo operators and small shops and a highway for larger teams, and it is valuable in both settings…
Reusable Software Components — Bottom-up design begins with known problems rather than the overall problem as a single starting point.
Revenue Recognition — A stage change alters what a company is allowed to do, including which early work can be booked as license revenue.
Reviewing and Testing Code — Writing code is often only about 20 percent of the problem; the other 80 percent is testing, reviewing, deploying, and…
Revisiting Neural Network Parameters — The total number of parameters in a neural network is the sum of the parameters in each layer.
Revocation Procedures for AI Agents — In regulated environments the security team examines the identity layer most closely, so an FDE needs a working answer for how…
Reward Baselines in Gradient Bandit Algorithms — Gradient bandit algorithms converge because their expected update is the gradient of expected reward.
Reward Baselines in Reinforcement Learning — Stochastic gradient ascent updates preferences only after reward feedback arrives.
Reward Prediction and Animal Learning — Classical conditioning is defined by the independence of the reinforcing stimulus from the animal's behavior.
Reward Prediction Error — The TD error and phasic dopamine activity can show matching patterns under a defined set of assumptions.
Reward Prediction Error Hypothesis — Dopamine neuron activity can shift from a reward to an earlier stimulus that predicts the reward.
Reward Prediction Error Hypothesis Explained — Dopamine neuron activity is related to reward prediction error, not simply to reward delivery.
Reward Signal — The reinforcement signal is the number that most directly guides changes in a policy or value function.
Reward-Modulated Spike-Timing-Dependent Plasticity — REINFORCE and error backpropagation can support comparable learning in multilayer networks, but they use different processes.
Reward-Modulated STDP — STDP makes synaptic change sensitive to the order and timing of neural spikes.
Reward-Modulated Synaptic Plasticity — Eligibility traces connect earlier neural activity with a later learning consequence.
Rewards and Penalties — A reward is an object or event that an animal will approach and work for.
Rewards and Penalties in Reinforcement Learning — Exploration tests unfamiliar actions so the agent can improve future decisions.
Rewards and Situations — Reinforcement learning asks how to map situations to actions to maximize a numerical reward signal.
Rewards in Reinforcement Learning — Return turns a sequence of rewards into one quantity using a specified function.
Ridge Regression — Feature manipulation transforms each original feature to create a resulting feature vector.
Risk and Prediction Error — Approximation error is a limitation in what the model can capture.
Risk Classification of AI Systems — The EU AI Act is Regulation 2024/1689, in force since August 1, 2024, with staged obligations through 2026 and 2027.
Risk Functions — The learning objective is to minimize the risk function L_D.
Risk of the Learned Hypothesis — Compare notions of learnability by asking what each definition is useful for.
Risk-Sensitive Optimization — Optimization-based artificial intelligence can behave unpredictably because a system may discover paths to its objective that…
RLM for convex-Lipschitz learning problems — Multiclass SVM handles several labels by scoring input-label pairs and selecting the largest score.
Role of Echo in a Startup — An Echo is a domain rebel: someone who came from the domain, understands it deeply, and has concluded that the current way of…
Role of Wrappers in AI Models — Anterograde amnesia in AI is a failure mode in which models do not learn from one session to the next and do not retain previous…
Root Mean-Squared Error — Batch updating keeps all episodes seen so far and repeatedly presents that growing batch until convergence.
Runnable Interface — A functional interface is a Java interface with a single abstract method, and Runnable is a standard example with one method, run.
Running a Java Program — The JVM is a virtual computer, not a physical machine.
Running a Java Program in Eclipse — A Java project in Eclipse is created from File, then New, then Java Project, where you choose a Project Name that can be anything.
Running Analysis Programs on Processed Data — The three-database pipeline consists of content.sqlite (raw, uncompressed), gmodel.py (transformation program), and index.sqlite…
Running Customer Discovery Interviews — A lead's role expands gradually, supported by a senior shadow who observes and gives specific feedback.
Running Deep-Learning Code in a Cloud Environment — A recent NVIDIA GPU is recommended for faster execution, not required for running the examples.
Running Java Programs From the Command Line — A jar file packs Java classes and resource files into one archive, and an executable jar names the class with the main() routine…
Running Python in Interactive Mode — The standard procedure for writing Python programs consists of four steps: open an editor, type code, save as a .py file, and…
Running Python Programs from the Command Line — A program written in a compiled language like C or C++ is converted from the source language i.e. C or C++ into a language that…
Running Python prompt on Windows — Using the Windows command line to run the Python interpreter.
Running Python Scripts — The interactive interpreter automatically prints expression results; scripts do not.
Running Your First Program — A program written in a compiled language like C or C++ is converted from the source language i.e. C or C++ into a language that…
Running Your First Python Program — Portability is Python's ability to run on many different operating systems without code changes, made possible by its…
Running Your First Web Spider — spdump.py reads spider.sqlite and displays pages in a five-element tuple format: (incoming_links, old_rank, new_rank, page_id…
Runtime and Optimization of Neural Networks — NP-hardness can arise in ERM implementation even for a network with one hidden layer and only four hidden neurons.