AUTOMATED DATA LABELING FOR SUPERVISED LEARNING MODELS

Abstract

The invention provides a system and method for automating the data labeling process for supervised learning models, addressing the time-consuming and costly challenge of manually labeling large datasets. The system utilizes pre-trained machine learning models, uncertainty estimation techniques, and active learning methods to generate initial labels for data. It then incorporates a human-in-the-loop mechanism to review and correct labels with low confidence, ensuring high-quality labeled data. This approach enhances efficiency, scalability, and accuracy in training machine learning models across various data types, including images, text, and sensor data, while minimizing human involvement and reducing costs.

Specification

Description:The embodiments of the present invention generally relates generally to the field of machine learning, and more specifically to a system and method for automating the process of data labeling for supervised learning models. Supervised learning, which requires large volumes of labeled data to train predictive models, often faces challenges related to the time, cost, and scalability of manual data labeling. This invention addresses these challenges by providing an automated approach to labeling data, making it more efficient, accurate, and scalable for use in a wide range of machine learning applications.

BACKGROUND OF THE INVENTION
The following description of related art is intended to provide background information pertaining to the field of the disclosure. This section may include certain aspects of the art that may be related to various features of the present disclosure. However, it should be appreciated that this section be used only to enhance the understanding of the reader with respect to the present , Claims:1. A method for automating data labeling for a supervised learning model, comprising:
receiving unlabeled data;
applying a pre-trained model to the unlabeled data to generate predicted labels;
evaluating the confidence of the predicted labels using an uncertainty estimation mechanism;
flagging data points with low confidence for human review;
receiving human input for flagged data points; and
incorporating the human input into the labeled dataset for training the supervised learning model.

2. The method of claim 1, wherein the uncertainty estimation mechanism comprises a confidence score associated with each predicted label.

3. The method of claim 1, wherein the human input is provided through a user interface that allows the human reviewer to correct the predicted label.

4. The method of claim 1, wherein the human input is incorporated into the training set through an active learning loop, wherein the model is iteratively improved based on human feedback.

Documents