Salesforce, inc. (20250068901). SYSTEMS AND METHODS FOR CONTROLLABLE DATA GENERATION FROM TEXT

SYSTEMS AND METHODS FOR CONTROLLABLE DATA GENERATION FROM TEXT

Organization Name

salesforce, inc.

Inventor(s)

Shiyu Wang of Atlanta GA (US)

Yihao Feng of Austin TX (US)

Tian Lan of Palo Alto CA (US)

Ning Yu of Palo Alto CA (US)

Yu Bai of Palo Alto CA (US)

Ran Xu of Mountain View CA (US)

Huan Wang of Palo Alto CA (US)

Caiming Xiong of Menlo Park CA (US)

Silvio Savarese of Palo Alto CA (US)

SYSTEMS AND METHODS FOR CONTROLLABLE DATA GENERATION FROM TEXT

This abstract first appeared for US patent application 20250068901 titled 'SYSTEMS AND METHODS FOR CONTROLLABLE DATA GENERATION FROM TEXT

Original Abstract Submitted

embodiments described herein provide a diffusion-based framework that is trained on a dataset with limited text labels, to generate a distribution of data samples in the dataset given a specific text description label. specifically, firstly, unlabeled data is used to train the diffusion model to generate a data distribution of data samples given a specific text description label. then text-labeled data samples are used to finetune the diffusion model to generate data distribution given a specific text description label, thus enhancing controllability of training.