Generative Subtask Creation Using User Task History
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Decomposing tasks into subtasks is time-intensive and requires significant user interaction with the system, leading to inconsistencies and varying complexities, which increases the time and difficulty in task completion.
Innovation Solution
A task management system that uses a generative model to automatically generate subtasks based on previously created tasks by the user, leveraging a hierarchy of tasks and supplemental examples to streamline the decomposition process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a user manually decomposes a task into subtasks, then the task can be completed by different users in parallel, but the process is time-intensive and requires frequent user interaction
Solution Approach 1:
The system performs preliminary action by automatically generating candidate subtasks using a generative model before the user needs to decompose the task. The model generates multiple candidate subtasks based on the task description, which are then presented to the user for selection or modification, significantly reducing the time and effort required for task decomposition
Solution Approach 2:
The system enables self-service by allowing the generative model to autonomously generate subtask candidates without requiring continuous user interaction. The model independently analyzes the task description and produces structured subtask candidates, reducing the burden on users and minimizing frequent interactions with the computer system
2Adaptability or versatility
If a user manually creates subtasks, then the task decomposition can be customized, but inconsistencies arise across multiple tasks
Solution Approach 1:
The system implements feedback by presenting generated candidate subtasks to the user for review, selection, and modification. User feedback on the candidate subtasks is used to refine and adjust the generated subtasks, ensuring they meet user expectations while maintaining consistency across different tasks through the standardized generation process
Solution Approach 2:
The system applies parameter changes by adjusting the generation parameters of the subtasks based on user preferences and historical data. The generative model learns from user interactions and modifies its output parameters to produce subtasks that are both consistent with user expectations and standardized across different tasks, resolving the inconsistency issue
3Ease of operation
If manual subtask creation is used, then user control over subtask details is maintained, but the complexity of managing subtasks increases
Solution Approach 1:
The system performs preliminary action by pre-generating multiple candidate subtasks with different levels of detail and structure before user review. This allows users to select from pre-prepared options rather than creating subtasks from scratch, maintaining user control while reducing the operational complexity of managing subtask creation
Data Source
AI summary
A task management system stores different types of content items, such as tasks, data objects, or interactions with tasks. Some tasks may be decomposed into subtasks that each correspond to a sub-action to complete so the action corresponding to a task can be completed. To more efficiently generate subtasks associated with a task, the task management system selects a set of supplemental tasks that were previously created by the user from whom the task was received. For example, the set of supplemental tasks includes tasks created by the user and in a common list as the task. Based on the task and the set of supplemental tasks, the task management system generates a prompt for a generative model. The generative model generates one or more subtasks for the task based on the set of supplemental tasks and information about the task.