Manage Tasks

πŸ“™ After a task is published, Task Admin can conduct task management actions on the My Tasks-My Tasks Management page. Team Admin/Owner can navigate to the Team Task Management tab to manage team tasks. Let's start to dive into the logic and system of managing tasks on BasicAI together!🧐

Task Management

For Task Admin, navigate to My Tasksand then click My Tasks Management. The page will display all the task information assigned to the Task Admin, along with single/bulk management tools.

For Team Admin/Owner, click the Team Task Management tab to manage all the team tasks.

πŸ“˜

My Tasks Management VS Team Task Management

The Team Task Management page displays a list of all team tasks for Team Admin/Owner. Team Admin/Owner can manage all team tasks here. Its display and functions are similar to My Tasks Management, but it has two additional operation options:

  1. For new tasks that have not been fully configured, team managers can view them on the Team Task Management page and click Configure to continue configuration or click Delete to delete them.
  2. In the bulk management toolbar, team managers can click Link to Project to bind multiple tasks to a project.

1. Task List Sheet

The task list sheet displays all of the task information and single-task operation buttons, such as:

Task ID, Task Name, Data type, Associated Dataset, Task Progress Bar, Task Status, Task Phase Details and Single-Task Operation Buttons (including Pause, Close, Copy, Delete)

πŸ’‘The βš™ button at the right corner of the task list sheet allows users to configure the table to be displayed with selected columns as shown below. After selecting certain columns, please click Apply.

2. Bulk Manage Tasks

Click the checkbox on the left side of each task, and then the Bulk Management Toolbar will display at the bottom of the page. This is a convenient way for Task Admin to manage several or all tasks at the same time.

Bulk Management ButtonsSub- ButtonsResults
Deselect All//
Task ActionsPausePausing tasks will lead to the loss of annotation/reviewing access for all workers.
ContinueResume the paused tasks, and then workers can continue working on the tasks.
DeleteAfter clicking Delete, a warning window will pop up, indicating that all comments, rejections, and other data related to the tasks will be removed. However, the original data source of the tasks will not be removed, and all annotated results that have been accepted will not be removed. Only new tasks in the configuration stage and tasks that have been closed can be deleted.
RestartRestart tasks, and only completed tasks can be restarted.
CloseAfter clicking Close, a warning window will pop up, indicating that all the tasks will not be able to restart and all unaccepted annotations will be discarded.
Bulk ConfigBasic SettingsClick Basic Settings, confirm the data type and annotation type of the tasks, and then jump to edit task basics settings
OntologiesClick Ontologies, and jump to edit ontology classes and classifications.
WorkflowsClick Workflows, and jump to edit task workflow.
QA RulesClick QA Rules, confirm the data type and annotation type of the tasks, and then jump to edit QA rules.

3. Task Info Summary

Clicking on any Task ID on the Task Management page will lead you to the task info summary page, which contains 5 tabs: Overview, Data, QA, Performance, and Setting.

3.1 Overview

The Task Overview page contains basic metrics and statistics for admins to track task progress and monitor the completion status of the task. The information on this page includes: Completed Data Counts, In-Acceptance Data Counts, Rejection Counts, Comment Counts, Task Progress Detail Bar, Data & Object Accuracy Diagram, and Stage Breakdown Bar Chart.

  • Progress: The Progress bar indicates the specific counts of data which are in the stage of Annotation, Review, Acceptance or Completion.
  • Accuracy: The Task Accuracy chart displays accuracy statistics in the dimensions of Data and Objects according to different data types (LiDAR Fusion or Image) and annotation types (Instance or Segmentation) of the task. In general, the indicators included in the chart (which might differ according to specific tasks) are as follows:
LiDAR FusionImage
InstanceLiDAR Data Accuracy, LiDAR Object AccuracyImage Data Accuracy, Image Object Accuracy
SegmentationLiDAR Data Accuracy, LiDAR Point AccuracyImage Data Accuracy, Image Object Accuracy

Only the data/objects with error comments will be defined as wrong data/objects when calculating LiDar Fusion Accuracy or Image Accuracy. The calculation methods of accuracy on BasicAI are:

By Data: Calculated in units of an image or LiDAR fusion data. Assuming that a user sets the data accuracy baseline to 95%, if there are 100 bounding boxes in the data, and less than 5 errors/comments, the data is considered accurate. Data Accuracy= accurate data counts/total data submitted counts.

By Objects: If a box is not marked with an error comment, it is considered accurate. Object Accuracy = 1 - object counts with error comments/total object submitted counts.

In addition, for Segmentation of LiDAR Data, LiDAR Point Accuracy= 1 - number of wrong Points/ total number of annotated Points.

  • Stage Breakdown: The Stage Breakdown bar chart provides adequate information about how the data in the task are distributed at every workflow stage and every status.

πŸ‘

Data Status Distributions

There are 6 statuses for data in a task, namely Not claimed, Suspended, Re-Review, Rejected, Working, and Submitted.

  1. Not Claimed: Data that have not yet been claimed by any worker in the task pool
  2. Suspended: Data that have been suspended by a worker
  3. Re-Review: When a reviewer receives the modified data from the same annotator again, the data will be in the Re-Review status
  4. Rejected: Data that have been rejected by others
  5. Working: Data that are currently being worked on by a worker during annotation, review or acceptance stage
  6. Submitted: Data that have been submitted by a worker

3.2 Data

The Data page allows admins to view all the data in the task, drill down to manage each piece of data, use useful attributes on the right sort&filter card to filter out problematic data, and then enter the data to add comments or re-assign it to other workers directly.

πŸ”– Data Card

Hover on the data card to preview, view in QA, and check information about the data. Click Preview annotations to see the annotation results on each data card if the data has been annotated. Task Admin and Inspector can see&click the view in QA button and enter the tool page to add comments to the previous worker's assignment.

πŸ”– Drop-down Select Box

The drop-down box at the top left corner supports users to select/unselect all the data and re-assign the data to the original worker or back to the task pool.

πŸ”– Sort&Filter Card

The Sort&Filter Card on the right side of the page allows users to sort and filter problematic data according to the provided attributes.

FunctionsOptionsRemarks
Sort Ascending by Item Name/
Descending by Item Name/
FilterAdded DateSelect data-added dates from xxx to xxx.
Assignee AnnotatorSelect the name of the assignee annotator.
Workflow StageSelect the stage of the data: annotation, review or acceptance.
Data StatusNot Claimed, Working, Re-review, Rejected, Suspended, Submitted. Please see the explanations for each data status in the 3.1 Overview section above.
ActivitiesWith Comments, With My Comments, Marked as Invalid (which means on the tool page, a worker marked the data as Invalid under the Validity tab), With Workflow Rejections (which means a reviewer rejected the data), With ACC Rejections (which means an inspector rejected the data), With Revise (Task Admin/Inspector revised the data), Paused for an Hour
Violation of RulesViolation of auto QA rules if configured and applied to the data before
With BatchSelect data with or without a batch.

3.3 QA

Please jump to the Automated QA guide for this part of information.

3.4 Performance

Please jump to the Performance guide for this part of information.

3.5 Setting

The configuration options on the 'Task-Setting' page are consistent with the ones when setting up or creating a new task, including Ontology-Class&Classification, Instructions, Workflow, Comments, Workflow Settings, and Annotation Settings. Using logic is also consistent. For more details, please refer to the task creation guide.

Shared Tasks

External Partner Admin who receives and accepts shared tasks can browse and manage tasks shared with them under the Shared Tasks section. The page information and operational logic are similar to 'Task Management'.