公开(公告)号：US20240428552A1

公开(公告)日：2024-12-26

申请号：US18821791

申请日：2024-08-30

Applicant: Lemon Inc.

Inventor： Yegor Klochkov ,  Yang Liu

IPC: G06V10/44 ,  G06T7/10 ,  G06V10/764

Abstract: There are proposed methods, devices, and computer program products for image processing. In the method, a reference dataset is obtained, the reference dataset comprising a plurality of reference samples, a reference sample in the plurality of reference samples comprising: a reference image and a reference label corresponding to the reference image, the reference label indicating a processing result of the image processing. An influence of the reference dataset on a loss is determined, the loss being used for updating an image model associated with the image processing based on the plurality reference samples. A hyperparameter is determined for updating the image model based on the influence of the reference dataset. The image model is updated based on the hyperparameter, the loss, and the plurality of reference samples. Therefore, the image model may be updated in an accurate and effective way.

公开(公告)号：US20250111263A1

公开(公告)日：2025-04-03

申请号：US18396627

申请日：2023-12-26

Applicant: Lemon Inc.

Inventor： Yegor Klochkov ,  Wenlong Chen ,  Yang Liu

IPC: G06N20/00

Abstract: The present disclosure describes techniques for balancing classification accuracy and fairness of a model trained to perform classification tasks. At least one bias score function corresponding to each sensitive attribute associated with instances classified by the model is configured. The at least one bias score function is configured to measure fairness on an instance level. At least one modification rule is generated based on the at least one bias score function and parameters. The at least one modification rule corresponds to at least one fairness criterion. The parameters are associated with a target level of the at least one fairness criterion. At least a subset of predictions are modified by applying the at least one modification rule to the predictions generated by the model. The modified predictions satisfy the target level of the at least one fairness criterion while maintaining the classification accuracy.

公开(公告)号：US20240419980A1

公开(公告)日：2024-12-19

申请号：US18821829

申请日：2024-08-30

Applicant: Lemon Inc.

Inventor： Yegor Klochkov ,  Yang Liu

IPC: G06N3/0985

Abstract: There are proposed methods, devices, and computer program products for language processing. In the method, a reference dataset is obtained, the reference dataset comprising a plurality of reference samples, a reference sample in the plurality of reference samples comprising: a reference text string and a reference label corresponding to the reference text string, the reference label indicating a processing result of the language processing. An influence of the reference dataset on a loss is determined, the loss being used for updating a language model associated with the language processing based on the plurality reference samples. A hyperparameter is determined for updating the language model based on the influence of the reference dataset. The language model is updated based on the hyperparameter, the loss, and the plurality of reference samples. Therefore, the language model may be updated in an accurate and effective way.