Nvidia patents method to create fake video data to test AI models
Nvidia describes a way to check if an AI model works well by looking at specific features in its data. If the model struggles with certain combinations of features, the company uses a generative AI to create new, synthetic video samples that include those specific features. These new samples are then used to test and improve the original model.
In the filing’s own words
A machine learning model (MLM) may be trained and evaluated. Attribute-based performance metrics may be analyzed to identify attributes for which the MLM is performing below a threshold when each are present in a sample. A generative neural network (GNN) may be used to generate samples including compositions of the attributes, and the samples may be used to augment the data used to train the MLM. This may be repeated until one or more criteria are satisfied. In various examples, a temporal sequence of data items, such as frames of a video, may be generated which may form samples of the data set. Sets of attribute values may be determined based on one or more temporal scenarios to be represented in the data set, and one or more GNNs may be used to generate the sequence to depict information corresponding to the attribute values.
What exactly they claim (claim 1)
1. A method comprising: determining one or more temporal patterns of one or more attributes in one or more temporal scenarios that correspond to one or more inference tasks; assigning, to a sequence of data items, values of the one or more attributes that vary to correspond to the one or more temporal patterns associated with the one or more temporal scenarios; generating, using one or more generative machine learning models (MLMs), a temporal data sample comprising the sequence of data items using the assigned values; and evaluating performance corresponding to at least one MLM and the one or more inference tasks based at least on applying the temporal data sample to the at least one MLM.
Facts
- Company
- Nvidia (NVIDIA Corporation)
- Official title
- DATA FOR MACHINE LEARNING MODELS
- Publication number
- US20260290075A1
- Application number
- 19677532
- Filed
- May 14, 2026
- Published
- Sep 24, 2026
- Inventors
- 6
- Claims
- 20
Classified under
G06V 40/176 Physics & computingG06V 10/62 Physics & computingG06V 10/774 Physics & computingG06V 10/82 Physics & computingG06V 40/164 Physics & computingMore from Nvidia in AI & software
- Nvidia patents AI coach that analyzes your gameplay to suggest strategies Oct 13, 2026
- Nvidia patents AI system that writes and voices character conversations Sep 10, 2026
- Nvidia patents system that uses AI to find bugs in game videos Sep 17, 2026
- Nvidia patents AI method to make digital characters move realistically Sep 8, 2026
- Nvidia patents AI system that builds image generators from text prompts Sep 15, 2026
Summaries are simplified from the filing’s own abstract and first claim. A patent filing is not a product announcement.
Patent data and drawings: U.S. Patent and Trademark Office (USPTO). Nebula is not affiliated with or endorsed by the USPTO.