3,135 articles
Unsupervised learning
Unsupervised learning, also called unsupervised machine learning, is a machine learning paradigm in which models learn patterns from unlabeled data, chiefly through clustering and dimensionality reduction.
Unsupervised representation learning
Unsupervised representation learning trains a model to convert unlabeled data into feature representations useful for later prediction tasks, without task-specific labels; self-supervised learning is its dominant modern form.
Update (SQL)
UPDATE, also written SQL Update, is the SQL statement that changes values in existing rows of a table, using SET for assignments and WHERE to select rows.
Upper confidence bound
The upper confidence bound (UCB) is a decision rule for multi-armed bandits and reinforcement learning that selects the action with the largest estimated reward plus an uncertainty bonus.
Upsampling
Upsampling is a signal and image processing technique that increases sampling rate or resolution by inserting new samples, computed by interpolation kernels or learned by transposed and sub-pixel convolution.
Upsert
Upsert (UPDATE or INSERT): a single atomic database statement that inserts a row if it does not exist and updates it if it does, keyed on a unique constraint; implemented as PostgreSQL ON CONFLICT, SQLite UPSERT, MySQL ON DUPLICATE KEY UPDATE, and MongoDB upsert:true.
Uptake (company)
Uptake is an American industrial artificial intelligence company whose machine learning software predicts failures in heavy equipment fleets, founded in 2014 by Brad Keywell and Eric Lefkofsky.
Utah Data Center
The Utah Data Center, code-named Bumblehive, is an NSA-led intelligence data storage facility at Camp Williams near Bluffdale, Utah, completed in 2014 for roughly $1.5 billion.
Vahan (vehicle registration database)
Vahan is India's national vehicle registration database, created by the National Informatics Centre and the road transport ministry to computerize records at regional transport offices.
Valkey
Valkey is an open source in-memory data store used as a database, cache, and message broker, forked from Redis in 2024 under the Linux Foundation.
Value iteration
Value iteration is a dynamic programming algorithm that computes the optimal value function and policy for a Markov decision process, first proposed by Richard Bellman in 1957.
Value learning
Value learning, central to AI alignment, is the problem of getting an AI system to learn what humans actually value from behavior and feedback, most practically through RLHF.
Vanishing gradient problem
The vanishing gradient problem is the shrinking of error gradients toward zero as they propagate backward through deep or recurrent neural networks, leaving early weights unable to learn.
Vapnik–Chervonenkis dimension
The Vapnik–Chervonenkis dimension (VC dimension) is the size of the largest set of points a classifier class can label in every possible way, central to statistical learning theory.
Variant Call Format
The Variant Call Format (VCF) is a standardized text format for storing DNA variants such as SNPs and structural variants, developed for the 1000 Genomes Project in 2011.
Variational neural annealing
Variational neural annealing (VNA) is an optimization method combining annealing with variational training of an autoregressive neural network, introduced in Nature Machine Intelligence in 2021.
Vector database
A vector database is a database system that stores, indexes, and searches high-dimensional vectors called embeddings, enabling similarity search for text, images, and retrieval-augmented generation.
Vector Institute (Canada)
The Vector Institute for Artificial Intelligence is a not-for-profit AI research institute in Toronto, launched in 2017 as one of three institutes under Canada's national AI strategy.
Vector quantization
Vector quantization is a lossy block coding technique that maps each input vector to the closest vector from a finite codebook, used for speech, image, and audio compression.
Vector space model
The vector space model is an information retrieval method representing documents and queries as term-weight vectors, ranked by cosine similarity, introduced by Salton in 1975.
Vector-quantized variational autoencoder
The vector-quantized variational autoencoder (VQ-VAE) is a generative model, introduced at DeepMind in 2017, that encodes data into discrete codebook tokens, avoiding the posterior collapse of conventional variational autoencoders.
Vectorized query execution
Vectorized query execution, also called vectorized or batch-at-a-time execution, processes columnar batches of about 1,000 to 4,096 values per operator call instead of one tuple, amortizing Volcano interpretation overhead; introduced in MonetDB/X100 (CIDR 2005), it underlies ClickHouse, DuckDB, and VectorWise.
Vehicle identification number
A vehicle identification number (VIN), also called a chassis number, is a unique 17-character code assigned to every road vehicle, encoding its manufacturer, attributes, and model year.
Vehicle routing problem
The vehicle routing problem (VRP), also called the truck dispatching problem, is an NP-hard combinatorial optimization problem that seeks minimum-cost routes for a fleet of vehicles serving customers.
Velox (query engine)
Velox is Meta's open source C++ database acceleration library, a composable execution engine used by Presto (Prestissimo), Spark via Apache Gluten, and AI/ML pipelines; it executes optimized plans with vectorized, columnar processing.
Vertica
Vertica, also known as HP Vertica, is a commercial analytic database that stores data in a columnar, massively parallel format, first released in 2005 and sold to Rocket Software in 2026.
Vertical federated learning
Vertical federated learning is a machine learning setting where parties holding different features for the same samples jointly train a model without sharing raw data.
VGGNet
VGGNet is a family of convolutional neural networks by Oxford's Karen Simonyan and Andrew Zisserman that reached 16 to 19 layers using only small 3×3 filters.
Video datasets for computer vision
A video dataset for computer vision is a curated corpus of labeled video clips for training models in action recognition and video understanding, from UCF101 to the Kinetics family.
View (SQL)
A view in SQL, also called a database view, is a named stored query used like a table. It simplifies queries, hides table structure, and restricts what users see.