Skip to content

Concept Glossary: E

This page lists English glossary entries for this letter. Entry bodies are assembled from term source files.

edge

  • Meaning: An edge is the connection or relationship unit between nodes in a graph.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: graph, node, weight
  • Core Section: P2-9.3
  • Appears in: P2-9.4, P4-14.1

backpropagation

  • Meaning: Backpropagation is the procedure that computes how much the loss at the output depends on each parameter by working backward through the network. More precisely, it applies the chain rule along the computation graph to distribute the final error signal across intermediate activations and weights.
  • Why it matters: A neural network cannot reduce loss unless it knows which direction each weight should move. Backpropagation provides that gradient signal efficiently, so learning becomes a concrete calculation rather than a vague idea that the network adjusts itself. It also explains why deep layers can be trained by tracing each connection's contribution to the final error.
  • Related concepts: gradient, loss function, chain rule
  • Core Section: P5-5.1
  • Appears in: P5-5.2, P5-6.1, P5-7.1

edge weight

  • Meaning: An edge weight is a value attached to an edge to represent distance, cost, strength, or importance.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: graph, weight, directed graph
  • Core Section: P2-9.3
  • Appears in: P2-9.4

embedding

  • Meaning: An embedding turns text, tokens, sentences, or documents into vectors that models and search systems can compare.
  • Why it matters: Embeddings are the basis for semantic comparison, vector search, and RAG, so they connect language data to computable similarity.
  • Related concepts: distributed representation, similarity search, word2vec, vector, representation
  • Core Section: P6-3.1
  • Appears in: P1-11.1, P1-13.1, P1-13.2, P1-13.3, P5-10.1, P6-3.2, P6-3.3, P6-3.4, P7-5.1

encoder

  • Meaning: An encoder reads an input and turns it into internal representations that reflect context.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: decoder, Encoder-Decoder, contextual representation
  • Core Section: P1-11.3
  • Appears in: P1-11.2, P6-20.1, P6-20.2

Encoder-Decoder

  • Meaning: An Encoder-Decoder separates reading an input sequence from generating an output sequence.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: Seq2Seq, fixed-length vector, Attention
  • Core Section: P1-11.2
  • Appears in: P1-11.3, P6-5.1

environment

  • Meaning: An environment is the outside world or interaction target that responds to an agent's actions with state changes and rewards.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: agent, state, reward
  • Core Section: P1-8.3
  • Appears in: P4-2.3

activation

  • Meaning: An activation is an intermediate value produced inside a model layer or computation step. It is not the final answer, but a temporary internal response created as input passes through the network.
  • Why it matters: Neural networks do not jump directly from input to answer. They build a chain of intermediate representations, and activations are the values that carry those representations forward. Understanding activations helps readers see a network as a sequence of computations rather than as a sealed answer box.
  • Related concepts: representation, vector, parameter
  • Core Section: P1-3.3
  • Appears in: P1-4.3, P1-8.2, P3-6.1, P3-6.2, P5-10.1, P5-10.2

activation function

  • Meaning: An activation function transforms a weighted sum or intermediate score before it is passed to the next layer or output. It introduces a non-linear response instead of passing a simple linear combination forward unchanged.
  • Why it matters: Repeating only linear combinations would not give a deep network much more expressive power. Activation functions let layers form more complex decision boundaries and representations, while also affecting gradient flow and training stability. They explain why deep networks need both stacked layers and non-linear response curves.
  • Related concepts: linear combination, activation, hidden layer
  • Core Section: P5-1.2
  • Appears in: P5-3.1, P5-3.2

environment variable

  • Meaning: An environment variable is a named setting value supplied from outside the program at execution time.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: redirection, PATH, runtime
  • Core Section: P2-7.8
  • Appears in: P2-7.9

epoch

  • Meaning: An epoch is one full pass through the entire training dataset.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: batch, training, loss curve, accuracy
  • Core Section: P5-6.2
  • Appears in: P2-13.2, P5-6.1, P7-4.1

error

  • Meaning: An error is the difference between an estimate or prediction and the actual value.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: estimation, loss, sample
  • Core Section: P2-5.3
  • Appears in: P2-5.4, P2-6.1, P2-6.3

error case

  • Meaning: An error case is a concrete failed or unstable case recorded so it can be traced again.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: review target, structural hypothesis, failure type, evaluation, retrospective
  • Core Section: P7-2.1
  • Appears in: P7-2.3, P7-3.2, P7-4.2, P7-summary

error sample

  • Meaning: An error sample is an individual sample whose prediction or execution result was wrong or unstable.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: wrong-case list, error case, sample-by-sample comparison table, evaluation record, review-target sample
  • Core Section: P7-3.1
  • Appears in: P7-3.2, P7-4.1, P7-4.2, P7-summary

estimation

  • Meaning: Estimation uses observed samples to infer a value or property of a broader population.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: population, sample, error
  • Core Section: P2-5.3
  • Appears in: P2-5.4, P2-6.1

evaluation

  • Meaning: Evaluation checks how well outputs meet answers, criteria, or comparison goals.
  • Why it matters: Evaluation defines what counts as good before scores are interpreted, so it is part of task design as well as measurement.
  • Related concepts: output, metric, task definition, validation, human evaluation
  • Core Section: P6-16.1
  • Appears in: P1-4.4, P1-12.3, P1-14.5, P1-14.6, P3-3.2, P3-8.2, P3-9.2, P6-16.2, P6-17.1, P6-17.2, P6-18.2, P7-2.1, P7-2.2, P7-4.2, P7-5.2, P7-7.2

evaluation mode

  • Meaning: Evaluation mode applies stable inference rules for measuring or serving a model rather than training it.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: training, training mode, test
  • Core Section: P5-6.4
  • Appears in: P5-8.2

evaluation record

  • Meaning: An evaluation record stores scores, failure samples, settings, notes, and next actions from an evaluation.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: evaluation, execution summary, wrong-case list, review summary, retrospective
  • Core Section: P7-3.2
  • Appears in: P7-4.2, P7-5.2, P7-7.2, P7-summary

event

  • Meaning: An event is a set of outcomes grouped by a condition of interest.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: outcome, sample space, probability
  • Core Section: P2-5.1
  • Appears in: P2-5.2, P2-5.3

evidence

  • Meaning: Evidence is information that supports, weakens, or refines a current judgment.
  • Why it matters: Evidence separates fluent output from supported output and makes review possible.
  • Related concepts: uncertainty, probability, probabilistic reasoning, provenance, review
  • Core Section: P1-12.3
  • Appears in: P1-6.2, P1-10.3, P7-5.1, P7-5.2

ANN, approximate nearest neighbor

  • Meaning: Approximate nearest neighbor search finds sufficiently close vector candidates quickly instead of always guaranteeing the exact closest item. It intentionally trades perfect optimality for speed at scale.
  • Why it matters: Large vector collections make exact comparison expensive. ANN methods explain why real search systems often prefer fast, good-enough candidates over exact answers that arrive too slowly. This concept connects vector search quality with latency, memory use, and service constraints.
  • Related concepts: nearest neighbor, search index, vector database
  • Core Section: P1-13.4
  • Appears in: P6-3.4

evidence review

  • Meaning: Evidence review checks whether sources actually support the current claim.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: review, evidence, provenance, factuality, recency
  • Core Section: P1-16.1
  • Appears in: P1-10.3, P1-12.3, P1-16.2, P7-5.1, P7-5.2

example

  • Meaning: An example is one observed case in data or one concrete case used for learning.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: data, label, input, sample
  • Core Section: P1-4.2
  • Appears in: P1-3.2, P1-12.1

exception

  • Meaning: An exception is a special case that does not fit an existing rule or general pattern.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: rule-based system, incomplete information, noise
  • Core Section: P1-6.1

excessive agency

  • Meaning: Excessive agency is a state where an agent or automation system has broader action authority than the task requires.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: least privilege, scope, approval, security, agent
  • Core Section: P1-15.3
  • Appears in: P1-10.3, P1-14.6, P7-6.2

execution environment

  • Meaning: An execution environment is the full set of conditions under which work runs, including tools, files, settings, and permissions.
  • Why it matters: The same model and prompt can behave differently depending on access, tools, and permissions, so environment is part of quality and safety.
  • Related concepts: harness, reproducibility, host
  • Core Section: P1-14.5
  • Appears in: P2-7.1

execution order

  • Meaning: Execution order records the actual order in which notebook cells or steps ran.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: notebook, hidden state, reproducible record
  • Core Section: P2-10.3
  • Appears in: P2-10.1, P2-10.2

execution scene

  • Meaning: An execution scene names the work moment a table, log, or summary belongs to.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: execution summary, comparison table, retrospective, run record, review
  • Core Section: P7-index
  • Appears in: P7-1.1, P7-2.1, P7-7.1, P7-summary

execution summary

  • Meaning: An execution summary condenses one run into its question, inputs, baseline, results, failures, and next review items.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: run record, baseline model, error case, review summary, retrospective
  • Core Section: P7-2.1
  • Appears in: P7-index, P7-3.1, P7-4.1, P7-5.1, P7-summary

  • Meaning: Exhaustive search checks every possible case or path without skipping candidates.
  • Why it matters: It gives a clear correctness baseline while showing why computation can become impractical.
  • Related concepts: search, search space, computational limit
  • Core Section: P1-7.1
  • Appears in: P1-7.2, P1-13.4

expert system

  • Meaning: An expert system is a rule-based AI system that encodes domain expert judgment as rules and knowledge.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: rule-based system, knowledge acquisition, inference engine
  • Core Section: P1-3.1
  • Appears in: P1-2.1, P1-2.2

explanation facility

  • Meaning: An explanation facility shows which facts and rules led to a conclusion.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: rule-based system, inference engine, fact
  • Core Section: P1-3.1
  • Appears in: P1-2.1

exploitation

  • Meaning: Exploitation chooses the action that currently looks best from information already gathered.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: exploration, policy, reward
  • Core Section: P1-8.3
  • Appears in: P4-2.3

exploration

  • Meaning: Exploration deliberately tries less-tested actions to gather more information.
  • Why it matters: Without exploration, a learner can miss better long-term actions; without exploitation, it may fail to use what it has learned.
  • Related concepts: exploitation, policy, reward
  • Core Section: P1-8.3
  • Appears in: P4-2.3, P4-19.3, P4-summary

expression

  • Meaning: An expression combines values, variables, operations, or function calls into a unit that can be evaluated.
  • Why it matters: This concept gives readers the vocabulary to separate roles, boundaries, and failure modes that otherwise get mixed together in AI systems and data work.
  • Related concepts: variable, function, output
  • Core Section: P2-2.1
  • Appears in: P2-3.1