Virtual Agent Task Flow Control via Probabilistic Graphical Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current dialog systems for virtual agents are inefficient and time-consuming due to a lack of customization for specific domains and individual users, resulting in a standard approach that does not effectively streamline task flow.
Innovation Solution
A system and method that uses artificial intelligence to organize task flow by receiving user utterances, identifying tasks, obtaining sets of rules, executing tasks, and running probabilistic graphical models to determine and suggest subsequent tasks, thereby customizing the task flow based on user interactions and preferences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a standard approach is used for designing virtual agent use cases, then the system is easier to implement and maintain, but the task flow efficiency and user experience are reduced due to lack of customization
Solution Approach 1:
The patent implements dynamic task flow by using probabilistic graphical models that adapt to user preferences and contextual information. The system transitions from static, predefined task sequences to dynamic, context-aware task recommendations that update based on user interactions and preferences, thereby improving task flow efficiency while maintaining implementation feasibility through structured probabilistic frameworks.
Solution Approach 2:
The system changes parameters by incorporating user preference data and contextual variables into the probabilistic graphical model. By adjusting the probability distributions and task selection parameters based on user feedback and behavioral data, the system achieves customization without requiring complete redesign of the virtual agent architecture, thus maintaining ease of implementation while improving efficiency.
2Device complexity
If a standard approach is used for designing virtual agent use cases, then the system complexity is reduced, but the adaptability to specific domains and individual users is worsened
Solution Approach 1:
The patent applies preliminary action by pre-defining the probabilistic graphical model structure and task templates that can be reused across different domains. This allows the system to maintain low complexity through standardized components while achieving high adaptability by configuring the pre-built model with domain-specific parameters and user preference data, eliminating the need to build custom systems for each domain.
Solution Approach 2:
The system achieves universality by designing a domain-agnostic probabilistic graphical model framework that can handle multiple domains and user types. The core architecture remains universal and reusable, while adaptability is achieved through configurable parameters, task templates, and user preference profiles that can be adjusted without changing the fundamental system structure, thus maintaining low complexity across diverse applications.
3Adaptability or versatility
If unlikely task suggestions are provided in the task flow, then more task options are available to users, but the task flow efficiency is reduced due to irrelevant suggestions
Solution Approach 1:
The patent implements feedback mechanisms by using probabilistic graphical models that continuously learn from user responses to task suggestions. The system monitors which suggestions users accept or reject and adjusts the probability distributions accordingly, providing increasingly relevant task options over time. This feedback loop ensures high task flow efficiency by filtering out unlikely suggestions while maintaining variety through context-aware recommendations.
Solution Approach 2:
The system applies partial action by providing a subset of relevant task suggestions rather than all possible tasks. The probabilistic graphical model calculates and presents only the most likely relevant tasks based on current context and user preferences, eliminating irrelevant options that would reduce efficiency. This selective approach maintains task option variety while improving efficiency by focusing on high-probability suggestions.
Data Source
AI summary
The present system and method may generally include organizing the task flow of a virtual agent in a way that is controlled by a set of rules and set of conditional probability distributions. The system and method may include receiving a user utterance including a first task, identifying the first task from the user utterance, and obtaining a set of rules related to the plurality of tasks. The set of rules may determine whether pre-tasks and/or pre-conditions are to be executed before executing the first task. The set of rules may also determine whether post-tasks and/or post-conditions are to be executed after executing the first task. The system and method may include executing the task; running a probabilistic graphical model on the plurality of tasks to determine a second task based on the first task; suggesting to the user the second task; and updating the probabilistic graphical model after a threshold number of runs.