Real-time Assisted Labeling

Real-time Assisted Labeling enable you to label your projects automatically without requiring configurations or AI knowledge. This extension simplifies the workflow, labeling speed, and accuracy.

Supported Labeling Types: Span Labeling

Introduction

Real-time Assisted Labeling is an intuitive feature designed to streamline the data labeling process in Datasaur, enabling users to efficiently label data without the complexities of multiple steps and configurations. Traditionally, users had to navigate through various settings to activate assisted labeling, which was time-consuming and inefficient. This feature addresses these challenges by providing seamless, real-time predictions that enhance both the speed and accuracy of the labeling process. By integrating advanced AI models, Real-time Assisted Labeling simplifies the user experience while maintaining high-quality outcomes.

Key Features

• Real-Time Predictions: Users receive immediate labeling suggestions as they work, allowing them to accept or reject predictions on-the-fly. This reduces the manual effort required and accelerates the labeling process.

• Simplified Workflow: The feature eliminates the need for extensive configurations. Users can initiate real-time labeling with minimal input, making the process more user-friendly.

• Batch Processing: The system predicts labels in batches, allowing users to review multiple suggested labels at once, thereby enhancing efficiency and reducing the time spent on labeling tasks.

Quick Start Guide

We are currently in the beta release phase. Please contact [email protected] to enable this feature.

Real-time Assisted Labeling by default is enabled on all license. However if it’s is disabled you can follow these step:

1. Go to Manage Extension and Enable Real-time Assisted Labeling.

2. The Real-time Assisted Labeling Extension should appear on the right side.

3. Make sure Automatically suggest labels is active

Setting up Real-time Assisted Labeling is straightforward. Follow these simple steps:

1. Initiate Real-Time Labeling: Begin labeling your data as usual. The system automatically prepares the Real-Time Assisted Labeling feature. You don't need to configure any settings.

2. Prediction: Once enough labels are provided, the system will begin suggesting predictions in real time.

3. Review Predictions: Keep an eye on the predicted labels that appear alongside your current line. Confirm or reject them as you go along, ensuring the quality of your dataset.

4. Finalize Labels: Once you’ve completed the labeling, review your results to ensure accuracy and consistency across the dataset.

With these easy steps, you can optimize your labeling workflow using the Real-time Assisted Labeling feature, enhancing both productivity and data accuracy.