Human-Machine Interface AI With Virtual Common-Sense Screening
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Human machine interface technologies often provide responses to user commands that lack common sense, leading to frustrating interactions and discouragement, as they fail to distinguish between appropriate and inappropriate commands, especially in contexts involving jest, sarcasm, or commands intended for novelty rather than practical use.
Innovation Solution
The integration of common sense reasoning into artificial intelligence systems through a virtual execution environment and 'software in the loop' and 'model in the loop' testing techniques, which evaluate potential responses in a simulated environment before executing them in the real world, using libraries of human interaction data to determine appropriate responses.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If artificial intelligence systems execute responses directly to user commands, then response speed and efficiency are improved, but the system generates inappropriate responses lacking common sense reasoning
Solution Approach 1:
The patent applies preliminary action by executing artificial intelligence in a virtual execution environment before deploying responses to the real world. The system pre-evaluates potential responses by having the AI actuate within a simulated environment, allowing common sense reasoning to be tested beforehand. This prevents inappropriate responses from reaching users while maintaining efficient processing through the virtual/real environment distinction.
Solution Approach 2:
The patent introduces a virtual execution environment as an intermediary between user commands and real-world responses. This intermediate layer acts as a mediator that evaluates whether AI-generated responses are appropriate before they are executed in the real world. The virtual environment serves as a buffer that filters out inappropriate responses while allowing suitable ones to pass through to the real execution environment.
2Reliability
If the system evaluates all potential responses before execution, then response appropriateness is improved, but processing time and system complexity increase
Solution Approach 1:
The system performs preliminary evaluation of AI responses in a virtual execution environment before real-world execution. This advance testing allows the system to identify and filter inappropriate responses without requiring extensive real-time processing when a user command is received, thus reducing the perceived processing time for appropriate responses.
Solution Approach 2:
The patent uses a virtual execution environment that replicates or copies the essential characteristics of the real execution environment. By creating this virtual copy, the system can evaluate responses in a simulated setting without consuming real-world resources or time, allowing rapid pre-screening of potential responses before actual execution.
3Ease of operation
If common sense reasoning is integrated into AI systems, then user-friendliness is improved, but device complexity increases
Solution Approach 1:
The patent segments the execution environment into distinct virtual and real components. Common sense reasoning capabilities are integrated into the virtual execution environment, separating the complexity of evaluating appropriateness from the simplicity of actual response execution. This segmentation allows the system to gain common sense reasoning capabilities without making the entire system overly complex, as the reasoning logic is confined to the virtual layer.
Solution Approach 2:
The virtual execution environment serves as an intermediary that handles the complexity of common sense reasoning. Rather than embedding complex reasoning logic throughout the entire system, the patent concentrates this functionality in the virtual environment, which mediates between simple user commands and appropriate real-world responses. This intermediary approach manages system complexity by localizing the sophisticated reasoning capabilities to a specific layer.
Data Source
AI summary
Methods, apparatus, systems and articles of manufacture are disclosed to add common sense to a human machine interface. Disclosed examples include a human machine interface system that having an actuator to cause artificial intelligence to execute in a virtual execution environment to generate a virtual response to a user input. The system also includes a virtual consequence evaluator to evaluate a virtual consequence that follows from the virtual response, the virtual consequence generated by executing a model of human interactions, and an output device controller to cause an output device to perform a non-virtual response to the user input when the virtual consequence evaluator evaluates the virtual consequence as positive.