Zefr, Inc.

Zerf, Inc.

Jonathan Morra

Oded Noy

Zachary McPherson

Ryan De Marigny

For generating policy labels, a method generates at least one characteristic question that represents an interest policy. The method identifies characteristic question examples from a searchable database, wherein at least one characteristic question example represents each answer to each characteristic question. The method generates first policy labels for each characteristic question of the characteristic question examples. The method generates a content set of search results from the searchable database for search terms generated from each characteristic question. The method generates characteristic labels for the content set using a classifier Multimodal Large Language Model (MLLM) and a confidence criteria. The method generates second policy labels from the characteristic labels for the content set. The method trains a classifier model based on the content set and the second policy labels. The method generates a given policy label for media content using the classifier model.

Training content labeling models

For predicting a concept prediction, a processor generates a review job that includes a review question and a corresponding expert concept prediction for an expert media set drawn from a media corpus. The processor generates media reviews from a plurality of reviewers using the review job for the expert media set. The processor extracts media features for the media of the media reviews. The processor trains a review model with the media reviews and corresponding media features. The processor predicts the concept prediction for media of the media corpus using the review model.

Growing community of inventors

El Segundo, CA, United States of America

Jonathan Morra