Robotics & Machine Learning Daily News2024,Issue(Feb.1) :127-131.

Researchers Submit Patent Application, 'User Authentication And Automatic Capture Of Training Data', for Approval (USPTO 20240020907)

Robotics & Machine Learning Daily News2024,Issue(Feb.1) :127-131.

Researchers Submit Patent Application, 'User Authentication And Automatic Capture Of Training Data', for Approval (USPTO 20240020907)

扫码查看

Abstract

From Washington, D.C., NewsRx journalists report that a patent application by the inventor CHOKSHI, Palak (San Ramon, CA, US), filed on July 15, 2022, was made available online on January 18, 2024. The patent’s assignee is Roblox Corporation (San Mateo, California, United States). News editors obtained the following quote from the background information supplied by the inventors: “Online platforms, such as virtual experience platforms and online gaming platforms, generally provide tools for creation, upload, and storage of 3D models. The 3D models may include a title and description text, for example, to help in identifying the content of the 3D models. While titles and descriptions are generally considered accurate, the text may be contextual or subjective in its description. “For example, a model for a chair and/or table may be described as a type of table or intended use (e.g., dining, breakfast, coffee, etc.). Such descriptions may be sufficient for human identification but may lack more generic labels or details that may be useful in machine learning and training. An artist or developer may sometimes intentionally mislabel a 3D model in order to circumvent automatic filtering of objectionable content (e.g., labeling a weapon as a tool, labeling a loot box as briefcase, etc.). Example embodiments are presented herein that overcome these and other drawbacks.

Key words

Artificial Intelligence/Business/Emerging Technologies/Machine Learning/Roblox Corporation

引用本文复制引用

出版年

2024
Robotics & Machine Learning Daily News

Robotics & Machine Learning Daily News

ISSN:
段落导航相关论文