Machine Learning

TensorRT

TensorRT is NVIDIA's software development kit (SDK) for high-performance deep learning inference on NVIDIA GPUs.

AI HardwareAI Tools & Products

Termination condition

A termination condition, also called a stopping criterion, convergence criterion, or halting condition, is a rule that decides when an iterative algorithm should stop running.

Training & Optimization

Test Set

A test set is a collection of examples reserved from model fitting and model selection so that it can evaluate a fixed machine learning model or a fully specified learning procedure.

Model Evaluation

Test loss

Test loss is the value of a loss function computed on a held-out test data set: data that was used neither for training nor for validation or model selection.

Test-Time Training (TTT)

Test-Time Training (TTT) is a family of machine learning techniques in which a model updates some of its own parameters at inference time

AI InferenceDeep Learning

Test-time compute

Test-time compute (also called inference-time compute scaling or test-time scaling) is the practice of allocating additional computation while a large language model answers a query

Artificial IntelligenceLarge Language Models

Text summarization

Text summarization is the natural language processing (NLP) task of automatically producing a shorter version of one or more documents that preserves the most important information from the original text.

Deep LearningNatural Language Processing

The Pile (dataset)

The Pile is an 825.18 GiB (approximately 886 GB) English text corpus designed for training large language models, assembled from 22 diverse, high-quality subsets spanning academic, professional, internet…

Data & DatasetsNatural Language Processing

The Stack v2

The Stack v2 is a large open dataset of source code released by BigCode in February 2024 as the training dataset behind the StarCoder2 family of code models.

Data & Datasets

Threshold (for decision trees)

In a decision tree, a threshold is the cut point used in an internal node's split test that decides which child subtree a sample is routed to.

TikTok

TikTok is a short-form video app owned by the Chinese technology company ByteDance, best known in artificial intelligence circles as the most widely studied production recommender system: its For You Page…

AI Policy & RegulationAI Tools & Products

Time Series

A time series is a sequence of observations recorded in time order, often at regular intervals: daily retail sales, hourly electricity demand, per-minute server latency, quarterly GDP.

Data ScienceDeep Learning

Time Series Analysis

Time series analysis is the statistical study of ordered observations when dependence across time is part of the signal rather than a nuisance to ignore.

Data ScienceStatistics

Timestep

A timestep is a discrete unit of time progression in a sequential process. The term shows up in many corners of machine learning and applied math, and it does not always mean the same thing.

Token

A token is the basic unit of text that a language model reads and writes: a word, a subword fragment, a single character, or a byte, produced by splitting text during a step called tokenization.

Deep LearningNatural Language Processing

Tokenization

Tokenization is the process of breaking text into smaller units called tokens, which serve as the fundamental input to natural language processing (NLP) systems and large language models (LLMs).

Large Language ModelsNatural Language Processing

Top-p and top-k sampling

Top-p sampling, also called nucleus sampling, is a text-generation decoding method for large language models (LLMs) that, at each step, samples the next token from the smallest set of highest-probability…

Large Language ModelsNatural Language Processing

Top-p sampling

Top-p sampling, also called nucleus sampling, is a stochastic decoding method for text generation in which the model samples from the smallest possible set of tokens whose cumulative probability mass exceeds a…

Large Language ModelsNatural Language Processing

TopK SAE

A TopK SAE (TopK sparse autoencoder) is a variant of sparse autoencoder that enforces sparsity by keeping only the K largest latent pre-activations for each input and zeroing all the rest

AI Safety

Topic model

A topic model is a statistical model that discovers the abstract "topics" hidden in a collection of documents, where each document is represented as a mixture of a small number of latent topics and each topic…

Natural Language ProcessingStatistics

Tower

In deep learning, a tower is a self-contained sub-network inside a larger model that encodes one specific input or feature group into an output, usually an embedding

Information RetrievalModel Architecture

Trackio

Trackio is an open-source experiment tracking library for Python developed by Hugging Face. It records metrics, configurations and media from machine learning training runs and displays them in a dashboard.

Developer ToolsMLOps

Training

Training in machine learning is the process of fitting a model's parameters to data so that the model can make accurate predictions or generate useful outputs.

Deep Learning

Training AI to Paint with Code

Training AI to Paint with Code is an experimental AI art project published by designer and researcher Surya Narreddi in March 2026.

AI ResearchArtificial Intelligence

Training Set

A training set is the portion of data used to fit a statistical or machine learning model.

Data & Datasets

Training compute

Training compute is the total number of floating-point operations (FLOP) performed to train a machine learning model, normally counted over the forward and backward passes of the final training run.

AI ComputeTraining & Optimization

Training run

A training run is a single, deliberate instance of training a neural network from scratch (or from a prior checkpoint) on a specified dataset, with a fixed compute budget, hardware allocation, and time horizon.

Training & Optimization

Training-Serving Skew

Training-serving skew is a difference between a machine learning model's performance during training and its performance during serving (production inference).

Data & DatasetsMLOps

Trajectory (Reinforcement Learning)

A trajectory in reinforcement learning is a sequence of states, actions, and rewards that an agent experiences while interacting with an environment.

Reinforcement Learning

Transfer Learning

Transfer learning is a family of machine learning methods in which information learned from one source domain or task is used while learning a target domain or task.

Deep Learning

Trigram

A trigram is a contiguous sequence of three items (most often three words) drawn from a sample of text or speech

Natural Language Processing

True negative

A true negative (TN) is a case that a binary classification model correctly predicts as belonging to the negative class: the true label is negative and the predicted label is also negative.

Model EvaluationStatistics

True positive

A true positive (TP) is a prediction that is correctly positive: the model predicts the positive class and the true label is also positive.

Model EvaluationStatistics

True positive rate (TPR)

The true positive rate (TPR) is the proportion of actual positive cases that a classifier correctly identifies as positive, computed as TPR = TP / (TP + FN), where TP is the number of true positives and FN the…

Model Evaluation

TruthfulQA

TruthfulQA is a benchmark designed to measure whether large language models (LLMs) generate truthful answers to questions.

AI BenchmarksAI Safety

Two-Tower Model

The two-tower model, also known as the dual encoder, bi-encoder, or Siamese network for retrieval, is a neural network architecture that encodes a query and a candidate item with two separate sub-networks…

Information RetrievalNeural Networks

U-Net

U-Net is a convolutional neural network architecture designed for biomedical image segmentation.

Computer VisionDeep Learning

UMAP (Uniform Manifold Approximation and Projection)

UMAP (Uniform Manifold Approximation and Projection) is a nonlinear dimensionality reduction technique that compresses high-dimensional data into a low-dimensional map (typically 2 or 3 dimensions) while…

Data Science

Unawareness (Fairness Through Unawareness)

Unawareness to a sensitive attribute, more commonly called fairness through unawareness (FTU), is a machine learning fairness approach that tries to make a model fair by simply not giving it the sensitive or…

AI Ethics

Underfitting

Underfitting occurs when a machine learning model is too simple to capture the underlying patterns in the data.

Undersampling

Undersampling is a class imbalance handling technique in machine learning that removes examples from the majority class of a training set so the minority class is no longer drowned out.

Statistics

Universal Approximation Theorem

The universal approximation theorem states that a feedforward neural network with a single hidden layer of finite width and a suitable nonlinear activation function can approximate any continuous function on a…

Universal Manipulation Interface

Universal Manipulation Interface (UMI) is an open-source system for collecting robot manipulation training data with a handheld, camera-equipped gripper instead of an actual robot.

Robotics

Unlabeled example

An unlabeled example is a data instance that has one or more features but no label, meaning it carries the inputs a model reads but not the target answer the model is meant to produce.

Unsupervised Machine Learning

Unsupervised machine learning is a type of machine learning that finds patterns, structures, and relationships in data that has no labels, with no human-provided answer key to learn from.

Unsupervised learning

Unsupervised learning is a branch of machine learning in which algorithms identify patterns, structures, and relationships in data without relying on labeled examples or explicit human guidance.

Artificial IntelligenceData Science

Uplift Modeling

Uplift modeling (also called incremental modeling, true lift modeling, or net modeling) is a set of machine learning and statistical techniques that predict the incremental impact of a treatment or action on…

Data ScienceStatistics

Upweighting

Upweighting is the practice of assigning a larger weight to certain training examples (or groups of examples) so they contribute more to the loss function and gradient updates than the rest of the data.

User matrix

In collaborative filtering and matrix factorization recommender systems, the user matrix (commonly written U or P) is the matrix of latent-factor vectors for users: each row is one user's embedding in a…

V-JEPA

V-JEPA (Video Joint Embedding Predictive Architecture) is a self-supervised video model from Meta AI that learns by predicting masked regions of a video in an abstract latent representation space rather than…

AI ModelsComputer Vision

VAPO (Value-based Augmented PPO)

VAPO (Value-based Augmented Proximal Policy Optimization) is a reinforcement learning framework for training large language models on long chain-of-thought reasoning tasks.

Reinforcement Learning

VC dimension

The Vapnik-Chervonenkis dimension, almost always shortened to VC dimension, measures the capacity of a class of binary classifiers by asking how many points the class can label in every possible way.

MathematicsStatistics

Validation

Validation in machine learning is the process of checking how well a trained model performs on data it did not see during training, using a held-out validation set to tune hyperparameters, choose between…

Validation Set

A validation set (also called a development set or dev set) is a subset of labeled data that is held out from the training set and used to evaluate a model's performance during development

Model Evaluation

Validation loss

Validation loss is the value of a model's loss function measured on a held-out validation set, data the model never sees during weight updates, and it is the primary signal practitioners use to judge how well…

Vanishing Gradient Problem

The vanishing gradient problem is a difficulty in training deep neural networks where the gradients used to update the network shrink exponentially as they are propagated backward through the layers, leaving…

Deep LearningNeural Networks

Variable importances

Variable importances, also called feature importances, are scores assigned to each input variable of a predictive model that measure how much that variable contributes to the model's output.

Interpretability

Variational Inference

Variational inference (VI), also called variational Bayes (VB), is a method in machine learning and statistics that approximates an intractable posterior distribution by turning Bayesian inference into an…

Statistics

Vector database

A vector database is a database that stores data as high-dimensional vectors (numerical embeddings produced by a machine learning model) and retrieves records by similarity rather than exact match

AI InfrastructureDeveloper Tools

Wasserstein Loss

Wasserstein loss is a loss function for training generative models that measures the distance between two probability distributions as the Wasserstein-1 distance

Generative AIMathematics