公开(公告)号：US20230186077A1

公开(公告)日：2023-06-15

申请号：US17841577

申请日：2022-06-15

Applicant: NVIDIA CORPORATION

Inventor： Hongxu YIN ,  Jan KAUTZ ,  Jose Manuel ALVAREZ LOPEZ ,  Arun MALLYA ,  Pavlo MOLCHANOV ,  Arash VAHDAT

IPC: G06N3/08 ,  G06N3/04

CPC classification number: G06N3/08 ,  G06N3/0481

Abstract: One embodiment of the present invention sets forth a technique for executing a transformer neural network. The technique includes computing a first set of halting scores for a first set of tokens that has been input into a first layer of the transformer neural network. The technique also includes determining that a first halting score included in the first set of halting scores exceeds a threshold value. The technique further includes in response to the first halting score exceeding the threshold value, causing a first token that is included in the first set of tokens and is associated with the first halting score not to be processed by one or more layers within the transformer neural network that are subsequent to the first layer.

公开(公告)号：US20220405583A1

公开(公告)日：2022-12-22

申请号：US17681632

申请日：2022-02-25

Applicant: NVIDIA CORPORATION

Inventor： Arash VAHDAT ,  Karsten KREIS ,  Jan KAUTZ

IPC: G06N3/08

Abstract: One embodiment of the present invention sets forth a technique for training a generative model. The technique includes converting a first data point included in a training dataset into a first set of values associated with a base distribution for a score-based generative model. The technique also includes performing one or more denoising operations via the score-based generative model to convert the first set of values into a first set of latent variable values associated with a latent space. The technique further includes performing one or more additional operations to convert the first set of latent variable values into a second data point. Finally, the technique includes computing one or more losses based on the first data point and the second data point and generating a trained generative model based on the one or more losses, wherein the trained generative model includes the score-based generative model.

公开(公告)号：US20250061612A1

公开(公告)日：2025-02-20

申请号：US18585286

申请日：2024-02-23

Applicant: NVIDIA Corporation

Inventor： Karsten Julian KREIS ,  Arash VAHDAT ,  Yilun XU

IPC: G06T11/00 ,  G06F3/14

Abstract: In various examples, systems and methods are disclosed relating to neural networks for synthetic data generation with discrete and continuous variable features. In training, an encoder can determine a plurality of encodings from a plurality of samples of training data, and the continuous generative model can operate as a decoder that is conditioned on the plurality of encodings to generate an estimated output to update the encoder and the continuous generative model. The discrete generative model can be trained over the plurality of encodings to learn to generate discrete variables corresponding to the distribution of information represented by the training data. At runtime, the discrete generative model can be used to generate a discrete variable from an input prompt, and can provide the discrete variable to the continuous generative model for the continuous generative model to generate an output, such an image, conditioned on the discrete variable.

公开(公告)号：US20220101145A1

公开(公告)日：2022-03-31

申请号：US17357738

申请日：2021-06-24

Applicant: NVIDIA CORPORATION

Inventor： Arash VAHDAT ,  Karsten KREIS ,  Zhisheng XIAO ,  Jan KAUTZ

IPC: G06N3/08 ,  G06N3/04

Abstract: One embodiment sets forth a technique for creating a generative model. The technique includes generating a trained generative model with a first component that converts data points in the training dataset into latent variable values, a second component that learns a distribution of the latent variable values, and a third component that converts the latent variable values into output distributions. The technique also includes training an energy-based model to learn an energy function based on values sampled from a first distribution associated with the training dataset and values sampled from a second distribution during operation of the trained generative model. The technique further includes creating a joint model that includes one or more portions of the trained generative model and the energy-based model, and that applies energy values from the energy-based model to samples from the second distribution to produce additional values used to generate a new data point.

公开(公告)号：US20240161250A1

公开(公告)日：2024-05-16

申请号：US18485239

申请日：2023-10-11

Applicant: NVIDIA CORPORATION

Inventor： Yogesh BALAJI ,  Timo Oskari AILA ,  Miika AITTALA ,  Bryan CATANZARO ,  Xun HUANG ,  Tero Tapani KARRAS ,  Karsten KREIS ,  Samuli LAINE ,  Ming-Yu LIU ,  Seungjun NAH ,  Jiaming SONG ,  Arash VAHDAT ,  Qinsheng ZHANG

IPC: G06T5/00

CPC classification number: G06T5/002 ,  G06T2207/20081 ,  G06T2207/20084

Abstract: Techniques are disclosed herein for generating a content item. The techniques include performing one or more first denoising operations based on an input and a first machine learning model to generate a first content item, and performing one or more second denoising operations based on the input, the first content item, and a second machine learning model to generate a second content item, where the first machine learning model is trained to denoise content items having an amount of corruption within a first corruption range, the second machine learning model is trained to denoise content items having an amount of corruption within a second corruption range, and the second corruption range is lower than the first corruption range.

公开(公告)号：US20220101144A1

公开(公告)日：2022-03-31

申请号：US17211681

申请日：2021-03-24

Applicant: NVIDIA CORPORATION

Inventor： Arash VAHDAT ,  Jyoti ANEJA

IPC: G06N3/08 ,  G06N3/04

Abstract: One embodiment of the present invention sets forth a technique for creating a generative model. The technique includes performing one or more operations based on a plurality of training images to generate an encoder network and a prior network, wherein the encoder network converts each image in the training images into a set of visual attributes, and the prior network learns a distribution of the visual attributes across the training images. The technique also includes training one or more classifiers to distinguish between values for the visual attributes generated by the encoder network and values for the visual attributes selected from the distribution learned by the prior network. The technique further includes combining the prior network and the classifier(s) to produce a trained prior component that, in operation, produces one or more values for the visual attributes to generate a new image that is not in the training images.

公开(公告)号：US20220101121A1

公开(公告)日：2022-03-31

申请号：US17211687

申请日：2021-03-24

Applicant: NVIDIA CORPORATION

Inventor： Arash VAHDAT ,  Jyoti ANEJA

IPC: G06N3/08 ,  G06N3/04 ,  G06K9/00

Abstract: One embodiment of the present invention sets forth a technique for generating images (or other generative output). The technique includes determining one or more first values for a set of visual attributes included in a plurality of training images, wherein the set of visual attributes is encoded via a prior network. The technique also includes applying a reweighting factor to the first value(s) to generate one or more second values for the set of visual attributes, wherein the second value(s) represent the first value(s) shifted towards one or more third values for the set of visual attributes, wherein the one or more third values have been generated via an encoder network. The technique further includes performing one or more decoding operations on the second value(s) via a decoder network to generate a new image that is not included in the plurality of training images.

公开(公告)号：US20240111894A1

公开(公告)日：2024-04-04

申请号：US18164215

申请日：2023-02-03

Applicant: NVIDIA Corporation

Inventor： Karsten Julian KREIS ,  Tim DOCKHORN ,  Tianshi CAO ,  Arash VAHDAT

IPC: G06F21/62 ,  G06N3/0455

CPC classification number: G06F21/6245 ,  G06N3/0455

Abstract: In various examples, systems and methods are disclosed relating to differentially private generative machine learning models. Systems and methods are disclosed for configuring generative models using privacy criteria, such as differential privacy criteria. The systems and methods can generate outputs representing content using machine learning models, such as diffusion models, that are determined in ways that satisfy differential privacy criteria. The machine learning models can be determined by diffusing the same training data to multiple noise levels.

公开(公告)号：US20220398697A1

公开(公告)日：2022-12-15

申请号：US17681625

申请日：2022-02-25

Applicant: NVIDIA CORPORATION

Inventor： Arash VAHDAT ,  Karsten KREIS ,  Jan KAUTZ

IPC: G06T5/00 ,  G06N3/08

Abstract: One embodiment of the present invention sets forth a technique for generating data. The technique includes sampling from a first distribution associated with the score-based generative model to generate a first set of values. The technique also includes performing one or more denoising operations via the score-based generative model to convert the first set of values into a first set of latent variable values associated with a latent space. The technique further includes converting the first set of latent variable values into a generative output.

公开(公告)号：US20220101122A1

公开(公告)日：2022-03-31

申请号：US17357728

申请日：2021-06-24

Applicant: NVIDIA CORPORATION

Inventor： Arash VAHDAT ,  Karsten KREIS ,  Zhisheng XIAO ,  Jan KAUTZ

IPC: G06N3/08 ,  G06N3/04 ,  G06K9/00

Abstract: One embodiment of the present invention sets forth a technique for generating data using a generative model. The technique includes sampling from one or more distributions of one or more variables to generate a first set of values for the one or more variables, where the one or more distributions are used during operation of one or more portions of the generative model. The technique also includes applying one or more energy values generated via an energy-based model to the first set of values to produce a second set of values for the one or more variables. The technique further includes either outputting the set of second values as output data or performing one or more operations based on the second set of values to generate output data.