Definition
In machine learning, diffusion models, also known as diffusion-based generative models or score-based generative models, are a class of latent variable generative models. A diffusion model consists of two major components: the forward diffusion process, and the reverse sampling process. The goal of diffusion models is to learn a diffusion process for a given dataset, such that the process can generate new elements that are distributed similarly as the original dataset. A diffusion model models data as generated by a diffusion process, whereby a new datum performs a random walk with drift through the space of all possible data. A trained diffusion model can be sampled in many ways, with different efficiency and quality.
Related concepts
AAAI Conference on Artificial IntelligenceActive learning (machine learning)AlexNetAnomaly detectionApprenticeship learningArtificial intelligenceArtificial neural networkAssociation rule learningAttention (machine learning)AutoencoderAutomated machine learningAutomatic summarizationBIRCHBatch learningBayesian networkBias–variance tradeoffBoltzmann distributionBoltzmann machineBoosting (machine learning)Bootstrap aggregatingBrownian dynamicsBrownian motionCURE algorithmCanonical correlationCluster analysisCoefficient of determinationComputational learning theoryComputer visionConditional random fieldConditioning (probability)Conference on Neural Information Processing SystemsConfusion matrixContinuity equationContrastive Language-Image Pre-trainingConvolutional neural networkCrowdsourcingCurriculum learningDALL-EDBSCANData cleaningData miningDecision tree learningDeepDreamDeep learningDenoiseDensity estimationDiffusionDiffusion (disambiguation)Diffusion processDimensionality reductionDiscrete diffusion modelECML PKDDEcho state networkElectrochemical RAMEmpirical risk minimizationEnsemble learningEuler–Maruyama methodEvidence lower boundExpectation–maximization algorithmFactor analysisFeature engineeringFeature learningFeedforward neural networkFokker–Planck equationFoundation modelFuzzy clusteringGated recurrent unitGaussian noiseGaussian processGenerative adversarial networkGenerative modelGlossary of artificial intelligenceGrammar inductionGraphical modelHadamard product (matrices)Heun's methodHidden Markov modelHierarchical clusteringHuman-in-the-loopImageNetImage denoisingIndependent and identically distributed random variablesIndependent component analysisInpaintingIntegration by partsInternational Conference on Learning RepresentationsInternational Conference on Machine LearningInternational Joint Conference on Artificial IntelligenceIsolation forestItô processJournal of Machine Learning ResearchK-means clusteringK-nearest neighbors algorithmKernel machinesLTX (AI Model)Lanczos resamplingLangevin equationLaplace operatorLatent diffusion modelLatent variable modelLeNetLearning curve (machine learning)Learning to rankLightricksLinear discriminant analysisLinear multistep methodLinear regressionList of datasets for machine-learning researchList of datasets in computer vision and image processingLlama (language model)Local outlier factorLogistic regressionLong short-term memoryMachine Learning (journal)Machine learningMamba (deep learning architecture)Markov chainMask (computing)Maurice TweedieMaximum a posteriori estimationMaximum likelihood estimationMaxwell–Boltzmann distributionMean shiftMechanistic interpretabilityMemtransistorMeta-learning (computer science)Ming-Hsuan YangMixture of expertsMulti-agent reinforcement learningMultimodal learningNaive Bayes classifierNatural language generationNatural language processingNeural fieldNeural network (machine learning)Neural radiance fieldNeuro-symbolic AINeuromorphic engineeringNoisy channel modelNon-equilibrium thermodynamicsNon-negative matrix factorizationNormal distributionNumerical methods for ordinary differential equationsOPTICS algorithmOccam learningOnline machine learningOntology learningOpen sourceOrdinary differential equationOutline of machine learningPerceptronPhysics-informed neural networksPolicy gradient methodPrincipal component analysisProbably approximately correct learningPrompt engineeringProper generalized decompositionPushforward measureQ-learningQuantum machine learningRandom forestRandom sample consensusReceiver operating characteristicRecurrent neural networkRegression analysisReinforcement learningReinforcement learning from human feedbackRelevance vector machineReparameterization trickReservoir computingRestricted Boltzmann machineRule-based machine learningScore (statistics)Scoring ruleSelf-organizing mapSelf-play (reinforcement learning technique)Self-supervised learningSemantic analysis (machine learning)Semi-supervised learningSergey LevineSigmoid functionSimulated annealingSora (text-to-video model)Sparse dictionary learningSpiking neural networkStable DiffusionStable distributionState–action–reward–state–actionStatistical classificationStatistical learning theoryStochastic differential equationStochastic gradient descentStructured predictionStyleGANSuper-resolutionSupervised learningSupport vector machineT-distributed stochastic neighbor embeddingT5 (language model)Temporal difference learningText-to-image modelThermodynamic betaTopological deep learningTransformer (deep learning architecture)Transformer (machine learning model)U-NetUnsupervised learningVapnik–Chervonenkis theoryVariational Bayesian methodsVariational autoencoderVariational inferenceVertical barVision transformerWasserstein metricWiener process
13 concepts already in your glossary