Intercept - An intercept is the constant term in a model's linear expression, added independently of the input features.
Interjection - An interjection is a word or short expression commonly used to convey a reaction, attract attention, greet someone, or provide conversational feedback...
Interlingual machine translation - Interlingual machine translation translates through a shared intermediate representation intended to capture relevant meaning independently of a particular...
Intonation phrases - Intonation phrases are stretches of speech grouped together by prosodic organization, such as a coherent pitch pattern and cues to a boundary.
IPA - IPA, the International Phonetic Alphabet, is a system of symbols and marks for representing speech sounds.
Kleene + - Kleene +, also called positive closure, is the operation that allows one or more repetitions of an expression or members of a language.
Kleene star - The Kleene star is an operation on a language or pattern that allows zero or more finite repetitions.
Knowledge citations - Knowledge citations are references attached to factual claims in a generated answer so readers can inspect the supporting sources.
Knowledge extraction - Knowledge extraction is the process of turning information in source material into an explicit, structured representation that can be stored, connected...
KV cache - A KV cache, or key - value cache, stores previously computed attention keys and values during autoregressive transformer inference.
Labiodental consonants - Labiodental consonants are speech sounds produced with a constriction involving the lower lip and upper front teeth.
Language engineering - Language engineering is the practical discipline of designing, building, evaluating, and maintaining systems that process human language.
Language resource - A language resource is an organized collection of language or speech data, or machine-readable descriptions of language, that people can use in language...
Latent semantic analysis - Latent semantic analysis, or LSA, represents relationships among terms and documents by reducing a term - document matrix to a smaller set of numerical...
Latent semantic mapping - Latent semantic mapping, or LSM, is a framework for representing discrete items as vectors in a smaller continuous space, using patterns across a collection...
Lateral sound - A lateral sound is a speech sound produced with airflow passing around one or both sides of the tongue rather than through an open central channel.
Lemma - A lemma is the canonical or dictionary-style form used to represent a set of related word forms.
Lesk algorithm - The Lesk algorithm is a dictionary-based approach to word-sense disambiguation: choosing which meaning of an ambiguous word fits its context.
Lexical choice - Lexical choice is the selection of words or lexical expressions to convey an intended message.
Lexical density - Lexical density measures how much of a text consists of lexical, or content-bearing, material.
Lexical resource - A lexical resource is an organized source of information about words and lexical expressions.
Lexical substitution - Lexical substitution is the task of finding a replacement word or expression for a target word in context while preserving the intended meaning.
Lexico-syntactic pattern - A lexico-syntactic pattern is a template that combines particular words with grammatical structure to identify a recurring linguistic configuration.
Lexicon - A lexicon is a collection of words or other lexical units together with information about them.
Light verbs - Light verbs are verbs that contribute relatively little of a construction's specific event meaning, while another expression, often a noun, supplies much of...
Linear chain CRF - A linear chain CRF is a conditional random field that assigns probabilities to sequences of labels using a chain of dependencies between neighboring label...
Linearly separable - Two labeled groups are linearly separable when a single linear decision boundary can place every member of one group on one side and every member of the...
Locative adverb - A locative adverb is an adverb that expresses where an action or situation occurs or, in a broad spatial use of the term, the direction associated with it.
Log odds ratio - A log odds ratio is the logarithm of the ratio between two odds. It compares the relative odds of the same event in two groups or conditions.
Log probabilities - Log probabilities are logarithms of probabilities.
Logic form - A logic form, in the NLP convention covered here, is a structured representation of a sentence using predicates and shared arguments.
Logistic function - The logistic function maps a real-valued input to an S-shaped output.
Logit - The logit of a probability p is its natural log odds: logit(p) = ln[p/(1 − p)], for 0 < p < 1.
Logit lens - The logit lens is an interpretability technique that maps intermediate language-model representations into vocabulary scores using the model's output readout.
Lookahead in regex - Lookahead in a regular expression checks whether a pattern does or does not match at the current position without consuming the characters inspected.
Loudness - Loudness is the perceived strength of a sound, from quiet to loud.
Low frame rate - Low frame rate in speech recognition means processing or emitting acoustic-model representations at fewer time steps per second than a reference configuration.
Machine-readable dictionary - A machine-readable dictionary is a dictionary represented electronically so that computer programs can access and process its contents.
Manner adverb - A manner adverb describes how an action or process happens. In “the technician worked carefully,” carefully specifies the manner of working.
Manner of articulation - Manner of articulation describes how the vocal tract shapes or restricts airflow when producing a speech sound.
Markov assumption - The Markov assumption says that the chosen present state contains the information from the past needed to predict the next state.
Mass nouns - Mass nouns are nouns used grammatically without presenting their referents as directly counted individual units.
Maxim, Gricean - A Gricean maxim is a conversational expectation used in H. P. Grice's account of how people interpret what others communicate beyond literal wording.
Maximum spanning tree - A maximum spanning tree is a tree that connects every vertex of a weighted graph while maximizing the sum of its selected edge weights.
Mel - A mel is a unit on a perceptually motivated pitch scale.
Mention detection - Mention detection is the task of identifying the text spans that refer to entities, or that count as mentions under a specified annotation scheme.
Mention-pair - Mention-pair is a coreference modeling approach that classifies a pair of textual mentions as referring, or not referring, to the same entity.