Concepts starting with L

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

Labeling Cases for Evaluation — An evaluation set is a collection of contracts with known correct answers, and it can grow during a project.
Labeling Interfaces for Evaluation Work — An evaluation set should be co-created with the customer, not written by the development team alone.
Labeling Pipelines in Validation — Validation is a separate phase from delivery. The source states that validation is not delivery practice and that it is the gate.
Lambda Expressions — A functional interface is a Java interface with a single abstract method, and Runnable is a standard example with one method, run.
Lambda Expressions and Variable Access — A local variable is effectively final if its value is never changed after it is initialized, and a parameter is effectively…
Lambda Expressions in Java — An ActionListener is a functional interface whose single method, actionPerformed, receives an ActionEvent, so a lambda…
Lambda Forms and List Comprehensions — A lambda statement creates an anonymous function object with one parameter and one expression.
Lambda Functions — List comprehensions are a concise syntax for creating a new list by transforming or filtering elements from an existing list.
Lambda Functions and the key Parameter — The DSU pattern transforms custom sorting into three simple, clear phases: decorate (add sort keys), sort (use Python's tuple…
Large Language Models — The source material for this article does not describe the cover artwork, so the article teaches the historical context given in…
Large Language Models and Their Applications — A transformer variant is a model built upon the original transformer architecture and adapted to a different family of tasks.
Large Margin Classification — Soft-SVM combines labeled training pairs and a positive λ parameter in an optimization problem.
Large State Spaces and Function Approximation — UCB can perform well on bandit problems, but that success does not make extension to general reinforcement learning…
Latent Representations — A latent space of images is a low-dimensional vector space whose points can be mapped to realistic-looking images.
Latent Space Sampling — GANs can generate new data by learning statistical structure from training data.
Launching an AWS EC2 GPU Instance — EC2 asks about creating new connection keys or reusing existing keys at the end of the launch process.
Layer Normalization — A TransformerBlock class builds its attention module, feed-forward network, two LayerNorm modules, and a dropout module from a…
Layers and Connections in Feedforward Networks — Activation functions determine how a neuron converts its scalar input into an output.
Layers and Models — A neural network's anatomy includes data representations, tensor operations, layers, models, and training mechanisms.
Layout Managers and Component Placement — The size of a component is not something to assume, because a layout manager often sets it and a user can resize the window.
Ldim — SOA begins with V_1 = H and maintains the hypotheses consistent with observed labels.
Leading Your First Small Engagement with Guidance — A senior shadow supports team members during a small engagement without doing the work for them.
Leaf Nodes and the Computation Graph — Calling .backward on the loss computes gradients for the leaf nodes of the graph, so you do not need to derive them by hand.
Learnability — Consistency evaluates whether a learning rule reaches a required performance level with high probability once the sample is…
Learnability of Bipartite Ranking Functions — Clustering groups similar data points into clusters.
Learnability of Convex Learning Problems — A learning problem in both definitions is written as (H, Z, ℓ).
Learning Agent and Environment — The reinforcement learning framework describes interaction between a learning agent and an environment.
Learning Algorithm — A prediction rule maps a domain point to a label.
Learning Halfspaces — Hinge loss is a convex surrogate for 0-1 loss.
Learning in Multilayer Neural Networks — REINFORCE and error backpropagation can support comparable learning in multilayer networks, but they use different processes.
Learning Mechanisms and Experimental Design — Classical versus instrumental conditioning is first a question about experimental setup.
Learning Models — Generalized loss functions provide a broad way to assign nonnegative values to model-domain pairs.
Learning Problems — X names the collection of objects being studied, while Y names the collection of possible labels.
Learning Problems and Hypothesis Classes — Linear regression is used to model relationships between explanatory variables and a real-valued outcome.
Learning Protocols — The learning protocol separates hypothesis construction from performance estimation.
Learning Rate Warmup — Cosine decay is a technique that modulates the learning rate throughout training epochs, making it follow a cosine curve after…
Learning State-Value Functions Under a Policy — Monte Carlo state-value estimation learns from complete observed returns rather than from a single outcome.
Learning Tasks and Hypothesis Classes — Computational complexity measures the number of operations an algorithm performs.
Learning Theory — The three references cover distinct areas of learning theory rather than one shared technical problem.
Learning to Predict and Learning to Control — Non-contingent eligibility traces do not correlate actions with consequent changes in the reward signal.
Learning with Neural Networks — A neural network architecture is specified by (V, E, σ), consisting of the graph and the activation function.
Least Squares — A polynomial predictor can represent relationships that require more than a straight-line predictor.
Least Squares Algorithm — Non-invertibility can arise when training instances do not span the entire space of R^d.
Lemma 27.4 — The chaining technique uses covering numbers to bound Rademacher complexity.
Limitations of Per-seat Pricing — Three structures replace per-seat pricing: milestone-based for engagement-shaped work, consumption-based for high-volume agent…
Limitations of the Discounted Setting — The average reward objective evaluates average reward over a policy's state distribution.
Linear Algebra — The lemma gives a probabilistic guarantee that distances between vectors can be preserved after projection into a…
Linear Algebra for Machine Learning — For linear regression with squared loss, least squares solves the empirical risk minimization problem.
Linear Algebra Review — Vectors in this setting are column vectors in finite dimensional Euclidean spaces.
Linear Classifiers in Binary Classification — VC dimension is a measure of classification capacity based on shattering.
Linear Combinations of Base Hypotheses — The capacity of L(B,T) is analyzed by counting dichotomies on a set C.
Linear Decision Boundaries — A halfspace hypothesis assigns one of two labels to an input vector.
Linear Dependence in R^d — VC dimension is a measure of classification capacity based on shattering.
Linear Function Approximation — A Fourier basis represents an approximation as a weighted sum of sine and cosine basis functions.
Linear Function Approximation Basics — SGD updates can be used with linear function approximation.
Linear Function Approximation in Reinforcement Learning — Sarsa controls the acrobot through repeated online interaction rather than through access to a simulator model.
Linear Functions — Logistic regression is used for classification tasks.
Linear Gradient-Descent Methods — Function approximation connects reinforcement learning value prediction with generalization across states and actions.
Linear MC Algorithm and Dutch Traces — Auxiliary vectors carry forward information from time steps before the final outcome is observed.
Linear Methods for Function Approximation — Linear TD(0) relies on the matrix A to control whether weight updates shrink or grow.
Linear Multiclass Predictors — The Natarajan dimension is a complexity measure for a hypothesis class.
Linear Predictors — Feature learning learns a function ψ that maps instances from X into d-dimensional feature vectors.
Linear Programming — Realizable ERM for homogeneous halfspaces can be expressed as a feasibility problem.
Linear Programming for Separable Classification — The Perceptron learns by correcting one currently misclassified sample at a time.
Linear Regression — Regression loss functions measure more than whether a prediction is wrong: they quantify the discrepancy between prediction and…
Linear Regression for Modeling Relationships — For linear regression with squared loss, least squares solves the empirical risk minimization problem.
Linear Search in Arrays — A phone directory built on an association list stores each name and number as a pair, so a lookup by key returns the associated…
Linear Separability — Compression keeps a smaller representation from which a zero-training-loss predictor can be reconstructed in the realizable case.
Linear Subspaces — PCA's optimization problem is a reconstruction problem: choose U and W to minimize total squared distance.
Linear Transformations — Dimensionality reduction maps vectors from R^d to R^n using a linear transformation, where n is smaller than d.
Link Analysis and Network Structure — PageRank is computed by running sprank.py, which iteratively calculates importance scores based on the link structure in your…
Linkage-Based Clustering — Clustering is an unsupervised organization problem rather than a labeled prediction problem.
Linkage-Based Clustering Algorithms — The supplied source pack does not define PCA or variance maximization.
Linked Lists and Their Operations — An abstract data type is a set of possible values and a set of operations on those values, specified without saying how the…
Linked Structures and Pointers — ArrayList and LinkedList implement the same List operations but store items differently: a growing contiguous array versus nodes…
Linking Tables with Foreign Keys — last_insert_rowid() is a built-in SQLite function that returns the ID of the most recently inserted row in your current database…
Lipschitz Coefficient — Higher dimension creates an exponential, not merely proportional, demand for training data in this analysis.
Lipschitz Continuity — 1-NN is analyzed as a sample-dependent hypothesis h_S for binary classification.
Lipschitz Functions — An r-cover provides a nearby representative for every point in A.
List and Dictionary Literals — There are four built-in data structures in Python list, tuple, dictionary and set . We will see how to use each of them and how…
List Comprehensions — List concatenation with + joins two or more lists into a single new list without modifying the originals.
List Comprehensions and Advanced List Methods — Lists are objects that bundle together data (the items they contain) and behavior (methods like append(), sort(), and…
List Indexing and Slicing — A list is an ordered, mutable collection of items created with square brackets and commas
List Methods and Modifications — Aliasing occurs when two or more variables refer to the same object in memory. The assignment b = a creates an alias, not a copy.
List Methods and Mutations — Lists are passed to functions by reference, meaning the function receives access to the original list, not a copy.
List Methods and Operations — List comprehensions are a concise syntax for creating a new list by transforming or filtering elements from an existing list.
List Methods and Sorting — Convert a dictionary to a list of (value, key) tuples so that sorting operates on the values (frequencies) rather than the keys…
List Methods for Adding and Removing Elements — Indices start at 0 and can be positive (counting from the start) or negative (counting from the end).
List Methods: append, extend, insert, and remove — Assignment creates a reference to a list, not a copy. Both variables point to the same object, so modifications through one…
List Methods: append, extend, pop, and del — Reassigning a function parameter (e.g., t = t[1:]) does not modify the original list passed to the function; it only changes…
List Methods: append, remove, and extend — Lists are mutable: you can change their contents after creation by reassigning elements using bracket notation on the left side…
List Methods: append, remove, and reverse — The sort() method arranges list elements in ascending order by modifying the list in place.
List Methods: append(), extend(), and insert() — List concatenation with + joins two or more lists into a single new list without modifying the originals.
List Methods: append(), insert(), remove() — Built-in functions (len, sum, max, min) extract information from lists without requiring custom loops.
List Methods: append(), remove(), and sort() — Built-in functions len(), max(), min(), and sum() enable efficient list analysis without manual loops.
List Methods: join() — A backup program automates file archiving by using Python to construct and execute a zip command with a timestamped filename.
List Methods: sort(), append(), and Other Modifiers — Aliasing means two variables point to the same list object, so changes through one variable affect the other.
List Models and Dynamic Lists — A JList is built from an array whose objects become the items the list displays.
List Mutability and In-Place Modifications — Index-based iteration combines range() and len() to generate indices, enabling you to read and modify list elements.
List Mutation and Methods — List slicing uses the syntax t[start:stop] to extract a portion of a list, where start is included and stop is excluded.
List Mutation and Side Effects — Identity (is) checks whether two variables refer to the same object in memory; equivalence (==) checks whether two objects have…
List Slicing and Copying — Python provides several methods for common list operations. Three of the most frequently used are append, extend, and sort. Each…
List Sorting and Ordering — Looping over a dictionary with for key in dict: iterates through the keys; use dict[key] to access the value.
List Sorting and the sorted() Function — Sorting dictionaries by value requires reconstructing tuples from (key, value) to (value, key) because Python's tuple sort…
Listener Interfaces in Java GUI Programming — An anonymous inner class is a class defined without a name, and a single new expression both declares it and creates an object…
Listing Files on the Command Line — The current directory, also called the working directory, is the folder whose files commands can reach by short name.
Lists and Collections — Counting uses a counter initialized to 0, incremented by 1 on each iteration, to track the number of elements processed.
Lists and Indexing — Parsing is the process of finding interesting lines in a file and extracting specific pieces of information from them.
Lists and Iteration — Sometimes we want to loop through a set of things such as a list of words, the lines in a file, or a list of numbers. When we…
Lists and List Operations — The three types of sequences mentioned above - lists, tuples and strings, also have a slicing operation which allows us to…
Lists and Mutability — Strings are immutable: once created, they cannot be changed. Attempting to assign a new value to a character at a specific index…
Lists and Sets in the Java Collection Framework — The Java Collection Framework divides generic data structures into collections, represented by Collection<T>, and maps…
Lists and Tuples — There are four built-in data structures in Python list, tuple, dictionary and set . We will see how to use each of them and how…
Lists vs. Dictionaries: Choosing the Right Data Structure — A dictionary is a mapping between keys and values, not a sequence indexed by position like a list.
Lists, Sets, and Priority Queues — A priority queue removes elements by priority, not by the order in which they were added.
Lists: Mutable Sequences — Unpacking is a syntactic feature that assigns multiple variables from a sequence in one statement, equivalent to index-based…
Literal Constants in Python — A literal constant is a value written directly in code that cannot be changed.
Littlestone Dimension — VC dimension measures the largest shattered set of instances.
Loading a Saved PyTorch Model — torch.save(model.state_dict(), file_name) writes a model's state dictionary to a .pth file, so the fine-tuned model can be…
Loading a Saved PyTorch State Dictionary — The recommended way to save a PyTorch model is to save its state_dict, passed as the first argument to torch.save.
Loading and Displaying Images From Files — An off-screen canvas is a BufferedImage, a subclass of Image, that stores a picture in memory instead of showing it directly.
Loading and Inspecting Instruction Datasets — The format_input function turns one dataset entry into a single prompt string that always includes an Instruction section.
Loading and Querying Geographic Data from SQLite — geodump.py converts database records into executable JavaScript code by extracting location names, latitudes, and longitudes and…
Loading GPT-2 Model Weights into GPTModel — Exercise 5.5 asks you to calculate two losses for a GPTModel with pretrained OpenAI weights, both on the 'The Verdict' dataset.
Loading Image Resources from Files — A Java Cursor combines an associated image with a hot spot, a Point that names the pixel matching the exact pointing position.
Loading Image Resources in Java — A Java cursor is a java.awt.Cursor object made of an associated image and a hot spot, which is a Point naming the pixel that…
Loading Images From Files in Java — Java has two fundamentally different kinds of Image: those read from external sources such as files or network connections, and…
Loading Images from Resource Files in Java — An icon in Java GUI is represented by an object of type Icon, and Icon is an interface rather than a class.
Loading Model Weights and Continuing Training — Adaptive optimizers such as AdamW store additional parameters for each model weight, and that data is separate from the weights…
Loading OpenAI GPT-2 Weights into a GPTModel Instance — Loading pretrained GPT-2 weights means calling load_weights_into_gpt with the gpt instance and the params, then moving the model…
Loading Pre-trained Model Weights — Pretraining is the stage where a model learns from text before being applied to a specific task, and this course treats it as…
Loading Pretrained GPT-2 Weights into GPTModel — A saved model can be resumed: load the model and optimizer in a new Python session or notebook, then continue pretraining for…
Loading Pretrained Models — Classification fine-tuning trains the model to output a correct class label instead of predicting the next token as in…
Loading Text Data into Python — Tokenization is the process of splitting input text into individual tokens, and it is a required preprocessing step for creating…
Local Classes and Variable Access — A local variable is effectively final if its value is never changed after it is initialized, and a parameter is effectively…
Local Variables and Scope in Java — A member variable is declared outside any subroutine but inside a class; a static member variable belongs to the class as a…
Logical and Physical Lines — A physical line is what appears in your editor; a logical line is what Python interprets as a single statement.
Logical Flow and Control Structures — Comparison operators test relationships between values and return True or False, forming the foundation of conditional logic in…
Logical Operators and Boolean Expressions — Short-circuit evaluation stops evaluating a logical expression once the result is already determined.
Logical Operators and Truth Tables — A boolean expression is any expression that evaluates to True or False; a condition is a boolean expression inside a conditional…
Logical Operators: and, or, not — Most statements (logical lines) that you write will contain expressions . A simple example of an expression is 2 + 3 . An…
Logistic Loss Function — Convexity of the logistic loss function with respect to w is the central property in this topic.
Long-Run State Distributions — Ergodicity separates temporary starting effects from long-run behavior.
Long-running agent harnesses — Agent readiness is a measure of how many deterministic validation loops are present inside a codebase.
Lookup Tables and Data Normalization — A junction table with a role attribute stores not just the fact of a relationship, but also the nature of that relationship…
Loop Conditions: Writing Expressions That Control Repetition — An infinite loop never terminates because its condition never becomes false.
Loop Control and Break Statements — Buffered reading prevents memory exhaustion by reading large files in fixed-size blocks and writing each block to disk before…
Loop Control Flow — When Python encounters a continue statement, it immediately stops executing the current iteration and returns to the top of the…
Loop Control Statements — Read a for loop as a cycle rather than as a single line.
Loop Control Variables — A for-each loop, written for ( T x : coll ), visits every object in a collection, and it can be used with any collection rather…
Loop Control with break and continue — We move the raw_input and if statements to inside the while loop and set the variable running to True before the while loop.…
Loop Control: Break and Continue — A countdown loop repeatedly decreases a control variable until a condition is no longer true, printing or using the variable at…
Loop Control: break and continue Statements — A for loop is a definite loop that processes each item in a known collection exactly once.
Loop else Blocks — We move the raw_input and if statements to inside the while loop and set the variable running to True before the while loop.…
Loop Exit Conditions — break; is Java's statement for jumping out of a loop.
Loop Idioms and Patterns — A for loop over a dictionary iterates through keys; to filter by value, use the index operator to retrieve each value and test…
Loop Invariants and Class Invariants — Both source methods use a while loop that visits each array index once, controlled by a counter i compared against A.length.
Loop Patterns and Best Practices — Python while loops can have an else clause, unlike C/C++ loops.
Loop Statements — A while loop is a control structure for condition-based repetition.
Loop Structures: For and While Loops — Iteration is the repeated execution of statements in a loop or recursive function, and understanding loop terminology is…
Loop Variable Initialization and Update — A while loop evaluates a boolean expression before each execution of its body.
Looping Over Array Elements — A parameter may have an array type such as int[], so one argument can deliver an entire array to a subroutine.
Looping Over Arrays with For Loops — Whole-array problems are solved by visiting every element with a for loop and combining what you find.
Looping with while Loops — A for loop executes a block of code once for each element in a list, with the loop variable taking on each element's value in…
Loops and Conditional Logic — The modulus operator (%) returns the remainder when one integer is divided by another.
Loops and Conditionals — File reading and transformation: Use a for loop to iterate through a file and apply string methods like upper() to each line.
Loops and Conditionals in Python — A grep simulator reads a file line by line, tests each line against a user-supplied regular expression, and counts how many…
Loops and Counters — Variable updates are not special operations; they follow the same evaluation rules as all assignments: Python evaluates the…
Loops and Iteration — The continue statement is used to tell Python to skip the rest of the statements in the current loop block and to continue to…
Loops and Iteration Patterns — Lists solve real-world data problems beyond simple numeric storage, including text processing, format parsing, and data analysis.
Loops and Program Correctness — A precondition is a condition that must be true before a piece of code runs, and it is used to ensure a program is correct.
Loops and Repeated Function Calls — The 3N+1 rule depends on one decision: if the current number is odd, the next number is 3 times the current number plus 1; if it…
Loops and User Input — The running total method maintains only a sum and count, discarding individual numbers after computing the average.
Loops: Iterating Through Files — find() returns the starting index of a substring if found, or -1 if not found
Loss Calculation and Training Functions — Instruction fine-tuning is a three-stage process: preparing the dataset, fine-tuning the LLM, and evaluating the LLM.
Loss Function Definition — Accuracy is the ratio of correct predictions to total examples, and calc_accuracy_loader tracks both as running counts across…
Loss Functions — A derivative explains how a small input change affects the output of a continuous function.
Loss Functions and Optimization — A machine learning project needs an observable definition of success before progress can be measured.
Loss Functions for Binary Classification — Manual gradient computation uses grad(loss, tensor) from torch.autograd to ask how the loss changes with respect to a chosen…
Loss Functions in Deep Learning — Style transfer assigns content and style to different input images and seeks a generated image that combines them.
LSTD Algorithm — Incremental updates keep LSTD from repeatedly rebuilding its accumulated quantities, but the matrix update remains O(n^2).
LSTM Networks and Sequence Memory — Character-level generation learns the relationship between a fixed-length character context and the character that follows it.