Diagnostic Assistance System for Shared Workload and AI Training
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing diagnostic systems require a substantial amount of time and labor to prepare training data, especially in fields like farm field diagnosis, where the growth cycle is long, leading to low inference accuracy and a heavy burden on customers or experts to investigate large datasets.
Innovation Solution
A diagnostic assistance system that enables multiple users to share the workload by storing and managing target data, allowing users to perform and share diagnostic tasks, with artificial intelligence learning from the results to provide automated diagnosis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a large amount of training data is prepared to train the learning model, then the inference accuracy is improved, but the time and labor required to prepare the training data increases substantially
Solution Approach 1:
The system performs preliminary actions by having multiple users manually diagnose a subset of target data before AI learning begins. This pre-prepared diagnostic data serves as initial training data, allowing the AI model to start learning with quality labeled data rather than requiring extensive data preparation time later. The users' diagnostic results are stored and used to train the AI, effectively performing the time-consuming labeling work in advance.
Solution Approach 2:
The system merges the diagnostic efforts of multiple users by collecting and aggregating their diagnostic results on the same target data. This consolidation creates a more robust training dataset faster than a single user could produce, combining multiple perspectives and improving data quality while reducing the total time needed compared to one person creating all training data independently.
2Measurement precision
If a large amount of training data is prepared to train the learning model, then the inference accuracy is improved, but the labor burden on users increases substantially
Solution Approach 1:
The system segments the diagnostic task by dividing target data into multiple portions and assigning different subsets to different users. This segmentation distributes the labor burden across multiple users, making the task more manageable for each individual while still accumulating sufficient training data overall. Each user only needs to diagnose a fraction of the total data rather than the entire dataset.
Solution Approach 2:
The system creates a multi-functional workflow where users perform diagnostic tasks that serve dual purposes: (1) providing diagnostic results for immediate use, and (2) generating training data for AI learning. This universality means the same user actions contribute to multiple goals, reducing the need for separate data collection efforts and making the overall process more efficient despite involving multiple users.
3Ease of operation
If multiple users perform diagnostic tasks and share results, then the workload on individuals is reduced, but the system complexity increases
Solution Approach 1:
The system introduces a server as an intermediary that manages the complexity of coordinating multiple users. The server handles data collection, storage, distribution of target data to users, aggregation of diagnostic results, and integration with the AI learning process. This centralized intermediary absorbs the system complexity, allowing individual users to interact with a simplified interface while the backend handles the complex multi-user coordination automatically.
Data Source
AI summary
A diagnostic assistance system for assisting a person in performing diagnosis of an object includes target data storage means for storing target data indicating a state of the object, diagnostic task means for providing each of a first user and a second user with the target data to enable each of the first user and the second user to perform a diagnostic task of the object, diagnostic result means for receiving a result of diagnosis of the object by each of the first user and the second user, and sharing means for enabling the first user and the second user to share result of diagnosis by each of the first user and the second user. The diagnostic task includes a first sub-task of investigating the target data to output a first diagnostic result and a second sub-task of using the first diagnostic result to output a second diagnostic result.