Cognitive Architecture Mask Variables for Embodied Agent Mode Switching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current artificial intelligence systems lack efficient and flexible mechanisms for cognitive mode-setting in embodied agents, which limits their ability to switch between different modes of behavior and connectivity, hindering their human-like capabilities.
Innovation Solution
The implementation of Mask Variables and Modulatory Variables in a Cognitive Architecture that allows for the modulation of connector strengths and switching between cognitive modes, enabling efficient and flexible control of behavior in embodied agents by overriding default connectivity and modulating information flow between modules.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional AI techniques use symbolic mental representations to guide behavior, then behavior control is achievable, but the system lacks flexibility in switching between different modes of connectivity and behavior
Solution Approach 1:
The patent implements dynamic connectivity by allowing connector strengths between modules to be modulated in real-time. Mask Variables and Modulatory Variables enable the system to dynamically switch between different cognitive modes (e.g., perception mode, action mode, rest mode) by adjusting information flow pathways, making the architecture adaptable rather than static.
Solution Approach 2:
The system changes operational parameters by using Mask Variables to override default connector strengths and Modulatory Variables to modulate information flow. This parameter-based control allows flexible transitions between cognitive modes without restructuring the entire architecture, resolving the contradiction between adaptability and complexity.
2Extent of automation
If subsumption architectures decompose behaviors into sub-behaviours organized in layers, then parallel processing is achieved, but the system lacks top-down control for cognitive mode-setting
Solution Approach 1:
The patent introduces Mask Variables and Modulatory Variables as intermediary control mechanisms between higher-level cognitive goals and lower-level sub-behaviours. These variables act as mediators that can override default connector strengths and modulate information flow, enabling top-down control for cognitive mode-setting while preserving the parallel processing benefits of subsumption architecture.
3Adaptability or versatility
If the system uses fixed connector strengths between modules, then computational efficiency is maintained, but the system cannot adapt to different cognitive contexts and tasks
Solution Approach 1:
The patent applies partial modulation by using Mask Variables to selectively override connector strengths only when cognitive mode transitions are required. Modulatory Variables provide graded control over information flow, allowing the system to maintain fixed connector strengths during stable cognitive modes (preserving computational efficiency) while enabling adaptive modulation when context changes require mode switching.
Data Source
AI summary
Embodiments described herein relate to a method of changing the connectivity of a Cognitive Architecture for animating an Embodied Agent, which may be a virtual object, digital entity, and/or robot, by applying Mask Variables to Connectors linking computational Modules. Mask Variables may turn Connectors on or off—or more flexibly, they may module the strength of Connectors. Operations which apply several Mask Variables at once put the Cognitive Architecture in different Cognitive Modes of behaviour.