United States
US 12,703,086 B2
Transformation for covariate shift of grasp neural networks
US 12,703,086 B2 is a granted patent from the United States Patent and Trademark Office (USPTO), dated 2026-08-11. Applicant: Siemens Aktiengesellschaft. Its PDF downloads from this page without an account, up to 5 a day.
- Applicant
- Siemens Aktiengesellschaft
- Inventors
- Juan L. Aparicio Ojea; Heiko Claussen; Ines Ugalde Diaz; Gokul Narayanan Sathya Narayanan; Eugen Solowjow; Chengtao Wen; Wei Xi Xia; Yash Shahapurkar; Shashank Tamaskar
- Publication date
- 2026-08-11
- Filing date
- 2023-10-17
- Document type
- B2 Granted patent
- Also written as
- US 12,703,086 B2 · US12703086 · 12703086
Abstract
A covariate shift generally refers to the change of the distribution of the input data (e.g., noise distribution) between the training and inference regimes. Such covariate shifts can degrade the performance grasping neural networks, and thus robotic grasping operations. As described herein, an output of a grasp neural network can be transformed, so as to determine appropriate locations on a given object for a robot or autonomous machine to grasp.
The opening of the abstract. The full text is in the PDF.
No account needed: 5 free downloads a day (5 left today). Batch downloads and Office Action citations need a free account.
Patent office website: United States Patent and Trademark Office (USPTO)
Look up another patent
Need many patents at once?
Paste a whole list of numbers. DownloadPatent finds the right version of each and packs the PDFs into one ZIP.
Sign up free