Invention Grant
- Patent Title: Supervised contrastive learning with multiple positive examples
-
Application No.: US17235992Application Date: 2021-04-21
-
Publication No.: US11347975B2Publication Date: 2022-05-31
- Inventor: Dilip Krishnan , Prannay Khosla , Piotr Teterwak , Aaron Yehuda Sarna , Aaron Joseph Maschinot , Ce Liu , Phillip John Isola , Yonglong Tian , Chen Wang
- Applicant: Google LLC
- Applicant Address: US CA Mountain View
- Assignee: Google LLC
- Current Assignee: Google LLC
- Current Assignee Address: US CA Mountain View
- Agency: Dority & Manning, P.A.
- Main IPC: G06K9/62
- IPC: G06K9/62 ; G06N3/08 ; G06V10/82

Abstract:
The present disclosure provides an improved training methodology that enables supervised contrastive learning to be simultaneously performed across multiple positive and negative training examples. In particular, example aspects of the present disclosure are directed to an improved, supervised version of the batch contrastive loss, which has been shown to be very effective at learning powerful representations in the self-supervised setting. Thus, the proposed techniques adapt contrastive learning to the fully supervised setting and also enable learning to occur simultaneously across multiple positive examples.
Public/Granted literature
- US20210326660A1 Supervised Contrastive Learning with Multiple Positive Examples Public/Granted day:2021-10-21
Information query