Concepts starting with N

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

n-step action-value backups — n-step Q(σ) unifies Sarsa and tree backup by allowing sampling and expectation to be selected across backup steps.
n-step Action-Value Backups with Q(σ) — ε-greedy behavior is defined relative to the current Q-values of the target policy.
n-Step Backups — The full return is the complete discounted reward sequence through episode termination.
n-Step Bootstrapping — The differential n-step return is G(n)t = R̄ + V̄(S(t+n)) − V̄(S(t)).
n-step Expected Sarsa — n-step Q(σ) is a unifying framework for action-value backups.
n-step Methods — Off-policy learning separates the policy collecting experience from the policy being learned.
n-Step Methods for Reinforcement Learning — n-step semi-gradient Sarsa joins a semi-gradient Sarsa control method with n-step bootstrapping.
n-step Returns and the λ-return — The forward view of eligibility traces connects eligibility-trace ideas with n-step returns and the λ-return.
n-step Sarsa — The n-step target is a quantity used by an action-value update rule.
n-Step Semi-Gradient Sarsa for Control — The algorithm separates the current environment time t from the delayed update time τ.
n-step state-value algorithms — Tree backup is an off-policy method distinguished by its lack of importance sampling.
n-step TD — The location of the importance sampling ratio depends on the type of update.
n-Step TD Algorithm and Error Reduction Property — The n-Step TD Algorithm waits for several steps of experience before updating an earlier state's value estimate.
n-Step TD Error — The differential n-step return is G(n)t = R̄ + V̄(S(t+n)) − V̄(S(t)).
n-step TD prediction — n-step Sarsa is the n-step version of Sarsa and an on-policy TD control method.
n-Step TD Prediction Methods — The full return is the complete discounted reward sequence through episode termination.
n-step Tree Backup — Expected action value is defined as a scalar random variable under the target policy.
n-Step Tree-Backup Algorithm — Episode-end handling requires two independent index checks.
Naive recurrent neural network implementation in NumPy — SimpleRNN is an actual Keras layer.
Namespaces and Scope — Fields are ordinary variables bound to class or object namespaces, scoped to their context and not global.
Naming Conventions and Best Practices — A variable is a name that refers to a value, allowing you to store and manipulate data in your program.
Naming Variables: Rules and Best Practices — Keywords are reserved words used by Python to define program structure and cannot be used as variable names.
Natarajan Dimension — The multiclass categorization goal is to learn h : X → [k].
Natarajan Dimension for Multiclass Classification — One-versus-All represents a multiclass hypothesis using one binary hypothesis for each label.
Natarajan's Lemma — The proof has separate lower-bound and upper-bound routes.
Native machine language — Just-in-time compilation translates Java bytecode into native machine language during program execution.
Natural-Language Processing — The course is divided into two parts.
Navigating and Querying XML Trees — XML is a formal language with syntax rules that go far beyond simple text searching. Comments, CDATA sections, entity…
Navigating BeautifulSoup Trees with find() and find_all() — BeautifulSoup transforms raw HTML strings into searchable tree-structured objects that mirror the document's logical hierarchy.
Navigating Common Scenarios in Customer Simulations — Bad news delivered with heavy softening can leave the customer unsure of the core fact and what they must do next.
Navigating Directories with cd — The current directory, also called the working directory, is the folder whose files commands can reach by short name.
Navigating Nested XML Structures — fromstring converts an XML string into a tree structure, enabling programmatic access to its data.
Navigating the DOM Tree with BeautifulSoup — Tag inspection in BeautifulSoup means accessing the individual components of a Tag object: its attributes (via get() or attrs)…
Navigating XML Trees with find and findall — findall retrieves a Python list of all XML nodes matching a specified path, enabling batch processing of similar structures.
Navigating XML Trees with findall() and XPath — ET.fromstring() converts an XML string into a tree structure of Element objects, making the hierarchical data queryable.
Nearest Neighbor Rule — The Nearest Neighbor rule shifts computational work to test time because it searches stored training examples when making a…
Nearest-Neighbor Selection — k-NN regression predicts a real-valued target by combining the targets of the k nearest neighbors.
Negative Indexing in Strings — A string is a sequence of individual characters that can be accessed one at a time using the bracket operator.
Nested Classes in Java — Compiling a class with a nested class creates two separate class files.
Nested Conditional Statements — A multiway branch built from if, else if, and else executes one and only one of its statements.
Nested Conditional Statements in Java — The dangling else problem occurs when nested if statements are interpreted differently by the compiler than the programmer…
Nested Conditionals — One conditional can also be nested within another. We could have written the three-branch example like this:
Nested Conditionals and Code Structure — Nested if statements can be flattened into a single conditional by using the 'and' operator to combine conditions.
Nested Conditionals and Complex Decision Logic — The and operator returns True only when both conditions are true; use it when you need multiple requirements to be satisfied…
Nested conditionals and when to use them — Chained conditionals use if, elif, and else to handle more than two possibilities in a clear, linear structure.
Nested Conditionals in Loops — The continue statement skips the rest of the current loop iteration and jumps back to the top of the loop to begin the next…
Nested Data Structures and Multi-Level Iteration — For loops are the most common and natural way to iterate through list elements, using identical syntax to string iteration.
Nested Dictionaries as an Alternative Structure — Composite keys are tuples used as dictionary keys to organize data by multiple related values simultaneously.
Nested Function Calls and the Call Stack — A function call redirects execution from the caller to the function body, then returns to the caller after the function completes.
Nested if Statements — You can have another if statement inside the if-block of an if statement and so on - this is called a nested if statement.
Nested List Comprehensions — Lambda functions used as sort keys provide a temporary comparison value for each element without modifying the original list.
Nested Lists and Multidimensional Structures — List concatenation with + joins two or more lists into a single new list without modifying the originals.
Nested Loops — You will see that we have two for loops. The outer loop is reading the lines of the file and the inner loop is iterating through…
Nested Loops and Complex Iterations — A for loop traverses a string by automatically assigning each character to a loop variable, one character per iteration.
Nested Loops and Iteration — urllib.request.urlopen() opens a web URL and returns a file-like object that you can iterate over line by line.
Nested Loops and Loop Control — The break statement exits a loop immediately, skipping any remaining code in the loop body.
Nested Loops for Hierarchical XML Data — findall() returns a Python list of all Element objects matching a given path in the XML tree.
Nested Loops in Python — You will see that we have two for loops. The outer loop is reading the lines of the file and the inner loop is iterating through…
Nested Loops: Combining Multiple Iterations — A for loop is a definite loop that processes each item in a known collection exactly once.
Netgame Hub and Connection Requests — The chat room application consists of two programs: a server, ChatRoomServer.java, and clients.
Network Client and Server Programs — Network input and output is blocking I/O, because those operations can block for indefinite periods of time, so network programs…
Network Communication and APIs — Unicode is a universal character encoding standard that assigns a unique code point to every character in every language…
Network Communication with Object Streams — Serializing an object means encoding it as a sequence of primitive-type data values so it can be output as bytes or characters…
Network Data Transmission — A socket is a two-way connection between two programs that allows both reading and writing of data, similar to a file but…
Network Game Communication — Many different network games share one requirement: an action taken by one player must be communicated over the network to the…
Network I/O Streams — InputStreamReader and OutputStreamWriter wrap an existing byte stream so that a program can read and write characters instead of…
Network Server Basics — A server that finishes one client before accepting the next, as the DateServer sample does, can be held up by a single slow or…
Networked Game Framework — The two-player Poker program follows a design similar to TicTacToe, with a Main class run by both players.
Networked Poker Game Architecture — The full Poker game state is known only to the PokerHub, not to the players.
Networked TicTacToe Game — The Hub is the single authoritative source for the TicTacToe game, and both network players connect to it.
Neural Machine Translation of Rare Words with Subword Units — Byte Pair Encoding is a tokenization method used in natural language processing, including in GPT-2.
Neural Mechanisms of Reward — Many drugs of abuse affect dopamine in striatal regions involved in normal reward-based learning.
Neural Network Approximation — The three references cover distinct areas of learning theory rather than one shared technical problem.
Neural Network Architecture Design — Hyperparameters are architecture-level choices, such as layer count, units or filters, activation functions, and dropout rates…
Neural Network Architectures and Activation Functions — Approximate ERM can remain computationally infeasible.
Neural network capacity — Reducing network size decreases neural network capacity and can mitigate overfitting.
Neural Network Loss Functions — Backpropagation calculates, rather than directly applies, the gradient of the loss with respect to network weights.
Neural Network Outputs — Each output of a multiple-output neural network may have its own loss function.
Neural Network Training — Analytical minimization identifies derivative-zero points, but it is not practical for large neural networks.
Neural Networks — Deep learning is a subset of machine learning.
Neural Networks for Classification and Regression — Part 1 is a staged introduction rather than a single isolated topic.
Neural Style Loss — Style transfer should be understood as retexturing guided by a style reference.
Neural Style Transfer — Variational autoencoders are generative models that learn compressed latent representations.
Neuromodulation and Reward Signals — Collective reinforcement learning changes the focus from one learner to the behavior of a population of reinforcement learning…
Neuromodulators and Brain Function — Dopamine is produced mainly by neurons in the SNpc and VTA of the midbrain.
Neuromodulatory Systems — The hedonistic neuron hypothesis applies reinforcement learning at the level of an individual neuron.
Neuron Function — A neuron has a typical three-part structure: cell body, dendrites, and axon.
Neuron Structure and Function — Dendrites receive information and the axon carries neuronal output.
Neuronal Plasticity — Eligibility traces preserve synapse-specific information about earlier activity.
Neuroscience and Collective Behavior — A hedonistic neuron is considered an individual reinforcement learning agent in Klopf's hypothesis.
Next-Word Prediction and Text Generation — GPT employs only the decoder portion of the original transformer and drops the encoder.
Next-Word Prediction in Language Models — An emergent behavior is an ability a model shows without having been explicitly trained to perform that task.
NN Rule — Higher dimension creates an exponential, not merely proportional, demand for training data in this analysis.
No-Free-Lunch Theorem — Shattering means realizing every possible 0/1 labeling on a chosen set.
Non-Blocking Sockets and select() — recv() returns the amount of data currently available in the network buffer, which may be less than the requested amount and…
Non-Markov State Signals — A state signal may combine immediate sensations with information derived from past sensations.
Non-Static Inner Classes — An inner class is a class defined inside another class, and its objects are always associated with an object of the containing…
Nonlinear activation functions — A sufficiently large single hidden layer with sigmoid units can approximate any continuous function on a compact input region to…
Nonlinear Function Approximation with Artificial Neural Networks — Universal approximation describes the expressive ability of one-hidden-layer ANNs, but it does not settle which architecture is…
Nonlinear Methods — Linear function approximation turns feature values and weights into a value estimate through a weighted sum.
Nonnegative Random Variables — Markov's inequality applies to nonnegative random variables.
Nonstationary Environments in Reinforcement Learning — Sample-average updating reduces the step size according to αn(a) = 1/n.
Nonstationary k-armed Bandit Task — Associative search requires both discovering good actions and linking them to the situations where they work.
Nonstationary Problems and Weighted Averages — The method gives greater weight to recent rewards and smaller weight to older rewards.
Nonstationary Reinforcement Learning Problems — UCB can perform well on bandit problems, but that success does not make extension to general reinforcement learning…
Nonstationary Target Functions — Backups provide training examples for function approximation in value prediction.
Nonuniform Learnability: Sample Size Flexibility — Nonuniform learnability makes sample-size requirements flexible across competing hypotheses.
Normalization in Neural Networks — Batch normalization normalizes across the batch dimension, pooling values across different inputs.
Normalizing Data: First, Second, and Third Normal Forms — A multi-table relational schema organizes data across separate tables to eliminate redundancy and improve data integrity. The…
Normalizing Database Schemas — A boolean flag on the User table cannot represent that the same person holds different roles in different courses.
Null References and Missing Objects — A variable of object type holds a reference to an object, never the object itself, so the object is not 'stored in' the variable.
Numbers in Python — Python has two main numeric types: int (integers) and float (floating-point numbers).
Numerical Timeseries Data — Timeseries data consists of measurements arranged in sequence at regular time intervals.
NumPy arrays — A tensor is a container for data, almost always numerical data.