Machine-Learned Device Settings from Sensor and User Override Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
User devices often expend computing resources on erroneously implemented settings due to sensor mischaracterizations, leading to user dissatisfaction and the need for manual overrides.
Innovation Solution
Implementing a machine learning model trained on user-controlled changes to automatically adjust device settings, reducing the likelihood of erroneous adjustments by learning from previous user interactions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automatic setting implementation is used based on sensor data, then productivity is improved, but reliability deteriorates due to sensor mischaracterizations causing erroneous settings
Solution Approach 1:
The system monitors user interactions with automatically implemented settings and feeds this information back to the machine learning model. When users manually override automatic settings, this correction data is used to retrain and improve the model, creating a continuous improvement loop that reduces erroneous settings over time
Solution Approach 2:
The machine learning model autonomously learns from user behavior patterns and automatically improves its setting recommendations without requiring explicit user programming. The system self-corrects by identifying patterns in user overrides and adjusting its predictions accordingly
2Reliability
If machine learning model is implemented to learn from user interactions, then reliability is improved, but device complexity increases
Solution Approach 1:
The machine learning model serves multiple functions: it predicts optimal settings, learns from user corrections, and continuously improves its accuracy. This single component handles both the initial setting recommendation and the ongoing refinement, reducing the need for separate complex systems
Solution Approach 2:
The system transitions from static sensor-based setting determination to dynamic machine learning-based prediction. The model parameters are continuously updated based on user interaction data, allowing the system to adapt to changing user preferences and environmental conditions without hardware modifications
3Reliability
If frequent manual overrides are required, then reliability is improved, but loss of time increases
Solution Approach 1:
The machine learning model performs preliminary learning during an initial training period by observing user interactions and preferences. This preliminary action builds a knowledge base that enables more accurate automatic settings from the start, reducing the need for subsequent manual overrides and time loss
Data Source
AI summary
In some aspects, a device may obtain first sensor data from a sensor configured to detect a characteristic associated with the device. The device may cause automatic implementation of a first setting for a feature of the device that is controllable by a user, the first setting based at least in part on the first sensor data. The device may detect a user-controlled change to the first setting for the feature. The device may obtain second sensor data from the sensor. The device may cause automatic implementation of a second setting, for the feature, that is identified by a machine learning model based at least in part on the second sensor data. The machine learning model may be trained to identify a setting for the feature based at least in part on information relating to the user-controlled change to the first setting. Numerous other aspects are described.