Lecture

Words, tokens, n-grams and Language Models

Graph Chat

Chat with Graph Search

Ask any question about EPFL courses, lectures, exercises, research, news, etc. or try the example questions below.

DISCLAIMER: The Graph Chatbot is not programmed to provide explicit or categorical answers to your questions. Rather, it transforms your questions into API requests that are distributed across the various IT services officially administered by EPFL. Its purpose is solely to collect and recommend relevant references to content that you can explore to help you answer your questions.

Description

This lecture covers the concepts of words, tokens, n-grams, and language models. It starts by discussing the ambiguity of defining words and tokens, then delves into n-gram models and their applications in language identification and spelling error correction. The lecture emphasizes the importance of understanding the probabilistic approach, including additive smoothing and Dirichlet priors. Key points include the challenges of out-of-vocabulary forms and the probabilistic approach to spelling error correction.

Official source

Ontological neighbourhood

Statistics

Statistical inference: Mathematical statistics

Related concepts (32)

Lexical analysis

Lexical tokenization is conversion of a text into (semantically or syntactically) meaningful lexical tokens belonging to categories defined by a "lexer" program. In case of a natural language, those categories include nouns, verbs, adjectives, punctuations etc. In case of a programming language, the categories include identifiers, operators, grouping symbols and data types. Lexical tokenization is not the same process as the probabilistic tokenization, used for large language model's data preprocessing, that encode text into numerical tokens, using byte pair encoding.

Categorical distribution

In probability theory and statistics, a categorical distribution (also called a generalized Bernoulli distribution, multinoulli distribution) is a discrete probability distribution that describes the possible results of a random variable that can take on one of K possible categories, with the probability of each category separately specified. There is no innate underlying ordering of these outcomes, but numerical labels are often attached for convenience in describing the distribution, (e.g. 1 to K).

Posterior predictive distribution

In Bayesian statistics, the posterior predictive distribution is the distribution of possible unobserved values conditional on the observed values. Given a set of N i.i.d. observations , a new value will be drawn from a distribution that depends on a parameter , where is the parameter space. It may seem tempting to plug in a single best estimate for , but this ignores uncertainty about , and because a source of uncertainty is ignored, the predictive distribution will be too narrow.

Dirichlet distribution

In probability and statistics, the Dirichlet distribution (after Peter Gustav Lejeune Dirichlet), often denoted , is a family of continuous multivariate probability distributions parameterized by a vector of positive reals. It is a multivariate generalization of the beta distribution, hence its alternative name of multivariate beta distribution (MBD). Dirichlet distributions are commonly used as prior distributions in Bayesian statistics, and in fact, the Dirichlet distribution is the conjugate prior of the categorical distribution and multinomial distribution.

Dirichlet process

In probability theory, Dirichlet processes (after the distribution associated with Peter Gustav Lejeune Dirichlet) are a family of stochastic processes whose realizations are probability distributions. In other words, a Dirichlet process is a probability distribution whose range is itself a set of probability distributions. It is often used in Bayesian inference to describe the prior knowledge about the distribution of random variables—how likely it is that the random variables are distributed according to one or another particular distribution.

Related lectures (64)

Topic Models

Introduces topic models, covering clustering, GMM, LDA, Dirichlet distribution, and variational inference.

Fundamental Limits of Gradient-Based Learning

Delves into the fundamental limits of gradient-based learning on neural networks, covering topics such as binomial theorem, exponential series, and moment-generating functions.

Gaussian Mixture Models: Data Classification

Explores denoising signals with Gaussian mixture models and EM algorithm, EMG signal analysis, and image segmentation using Markovian models.

Dirichlet-Multinomial Model

Discusses the Dirichlet distribution, Bayesian inference, posterior mean and variance, conjugate priors, and predictive distribution in the Dirichlet-Multinomial model.

Probability and Statistics