Automated Task Definition in Interactive Communication Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods lack an efficient way to automatically identify and partition sub-graphs representing tasks within interactive communications systems (ICS) from directed graphs, especially in complex systems like IVR menu systems, where manual task definitions are cumbersome and existing graph partitioning methods are not suited for finding sub-graphs of varying sizes.
Innovation Solution
A method and system that identify hub nodes with high in-degrees as unavailable for task consideration, compute scores for available subgraphs based on depth and size, and iteratively mark nodes as part of tasks, converting the identified hub nodes and subgraphs into an XML format for task definition, allowing user input for customization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If manual task definitions are used in complex ICS like IVR menu systems, then task definitions can be customized and reviewed, but the process becomes cumbersome and time-consuming
Solution Approach 1:
The system automatically identifies hub nodes and partitions the directed graph into connected subgraphs representing tasks without requiring manual intervention. The algorithm self-services by computing in-degrees, selecting hub nodes, and generating task definitions autonomously from the ICS flow graph.
Solution Approach 2:
The method extracts hub nodes (nodes with high in-degree) from the directed graph and uses them as seeds to identify connected subgraphs. By taking out these critical nodes first, the system efficiently partitions the graph into meaningful task components, automating what would otherwise require manual analysis.
2Productivity
If existing graph partitioning methods are used, then partitioning can be performed, but they are not suited for finding sub-graphs of varying sizes representing tasks
Solution Approach 1:
The algorithm treats different regions of the graph differently by using local properties (in-degree values) to identify hub nodes. Each connected subgraph is processed locally around its hub node, allowing the method to adapt to varying subgraph sizes and structures rather than applying a uniform partitioning approach throughout the entire graph.
Solution Approach 2:
The method changes the parameter used for partitioning from traditional graph theory metrics to in-degree values. By using in-degree as the key parameter for identifying hub nodes and seeds for connected subgraphs, the system adapts to the specific characteristics of ICS flows and can handle subgraphs of varying sizes effectively.
3Reliability
If all nodes in the directed graph are considered for task definition, then comprehensive coverage is achieved, but computational complexity increases
Solution Approach 1:
The directed graph is segmented into multiple connected subgraphs, each rooted at a hub node. This segmentation divides the computational problem into smaller, manageable pieces that can be processed independently. Each subgraph represents a specific task or functional area, reducing the overall computational complexity while maintaining comprehensive coverage of all nodes.
Solution Approach 2:
The algorithm performs preliminary identification of hub nodes based on in-degree calculations before proceeding to identify connected subgraphs. This preliminary action filters out nodes that are unlikely to be task starters (those with low in-degree), reducing the search space and computational complexity for subsequent subgraph identification while ensuring comprehensive coverage of relevant task areas.
Data Source
AI summary
Task analysis of an interactive communication system (ICS) can be performed manually. Manual task analysis is costly and time-consuming process. In an embodiment, a method of defining tasks within an ICS includes identifying hub node(s) to be marked as unavailable from consideration as nodes within a task. The at least one hub node can be within a directed graph representing flows through an ICS. The method further includes, from available nodes, automatically identifying a connected subgraph that corresponds to nodes representing an area of functionality defining a task within the ICS. The method additionally includes repeating the identifying of the connected subgraph at least one time. The method also includes outputting an indicator of the at least one hub node identified and the connected subgraphs that represent corresponding areas of functionality defining respective tasks. Therefore, task analysis is improved by extracting task definitions from the graph data automatically.