Smart Television Out-of-Tab Detection Using Local AI Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Smart devices, such as smart televisions, struggle to accurately determine when they are out-of-tab due to network connectivity issues, leading to inaccurate media crediting as they cannot differentiate between being off and losing connection, resulting in incomplete or inaccurate data transmission to servers.
Innovation Solution
Implementing an AI-based model trained on panelist data and smart television data to differentiate between in-tab and out-of-tab states by recognizing patterns and associations, using machine learning techniques like logistic regression and random forests to predict connectivity status based on interaction data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If smart devices transmit media exposure data to servers, then data accuracy is improved, but network connectivity requirements increase causing devices to go out-of-tab
Solution Approach 1:
The system performs preliminary actions by training AI models offline using historical panelist data and smart television data. The trained models are then deployed to devices, enabling them to make in-tab/out-of-tab determinations locally without requiring continuous network connectivity for model operations.
Solution Approach 2:
The patent introduces AI models as intermediaries between raw smart device data and tab status determination. These models process features like tuning data, disconnect data, and media exposure data to infer whether the device is in-tab or out-of-tab, bridging the gap between incomplete data and accurate status determination.
2Measurement precision
If smart devices continuously monitor tab status, then media crediting accuracy is improved, but device complexity increases
Solution Approach 1:
The monitoring system is segmented into modular components: data collection modules that gather tuning data, disconnect data, and media exposure data; feature extraction modules that prepare inputs; and AI model inference modules that determine tab status. This segmentation allows each component to be optimized independently and simplifies deployment.
Solution Approach 2:
The system implements self-service by enabling smart devices to autonomously determine their own tab status using locally deployed AI models. Devices process their own collected data and make independent determinations without requiring complex centralized monitoring infrastructure, reducing overall system complexity.
3Measurement precision
If AI models are trained on panelist data, then out-of-tab detection accuracy is improved, but data processing time increases
Solution Approach 1:
The system performs the computationally intensive model training action in advance using historical panelist data and smart television data. Once trained, the models are deployed to production environments where they perform rapid inference on new data, eliminating training time from the operational data processing pipeline.
Data Source
AI summary
Methods, apparatus, systems and articles of manufacture to determine whether a smart device is in-tab are disclosed. An example apparatus includes memory; instructions in the apparatus; and processor circuitry to execute the instructions to: provide smart television data from a smart television as an input to a model to generate an output, the smart television data being included in population data from a population of smart televisions; determine a tab status of the smart television based on the output; in response to the tab status of the smart television being out-of-tab, remove the smart television data from the population data; and credit media based on the population data.