ML-Based Data Object Routing in Process Flows
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing database systems fail to efficiently route data objects between process paths due to a lack of consideration for the underlying features of the data objects, leading to inefficient or ineffective data processing.
Innovation Solution
Implementing a machine learning model that uses data object features to determine the optimal path within a process flow, performing random routing to collect key performance indicators and update models dynamically, and allowing user input to address potential biases in the routing process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a single path is selected for process flow without considering data object features, then the process flow is simple to implement, but data processing efficiency and effectiveness deteriorate
Solution Approach 1:
The patent implements dynamic routing by using machine learning models that adapt to different data object features. The system transitions from static single-path routing to dynamic multi-path routing where the optimal path is selected based on real-time analysis of data object characteristics, allowing the routing decision to change according to the specific data being processed.
Solution Approach 2:
The patent changes the routing parameter from a fixed path selection to a feature-based dynamic selection. By introducing machine learning models that analyze data object features (such as data type, size, format) and map them to optimal paths, the system transforms the routing decision into a parameter-driven process that adapts to different data characteristics.
2Reliability
If multiple paths are evaluated using A/B testing, then path selection can be optimized, but the process becomes complex and may fail to account for underlying data object aspects
Solution Approach 1:
The patent replaces the mechanical A/B testing process with a machine learning-based routing system. Instead of manually evaluating multiple paths through statistical testing, the system uses trained ML models that automatically analyze data object features and predict optimal paths, substituting the manual evaluation mechanism with an automated intelligent system.
Solution Approach 2:
The machine learning model performs self-service by automatically analyzing data object features and making routing decisions without requiring manual intervention. The model continuously learns from data patterns and autonomously selects the optimal path based on the underlying aspects of the data objects, eliminating the need for complex manual A/B testing processes.
3Loss of information
If random routing is used to collect performance data, then model training data is gathered, but the routing is not optimized for individual data objects
Solution Approach 1:
The patent implements feedback mechanisms where the outcomes of data processing are used to train and improve the machine learning model. By collecting key performance indicators from both random routing and intelligent routing, the system creates a feedback loop that continuously refines the model's ability to predict optimal paths, using actual performance data to improve future routing decisions.
Data Source
AI summary
Methods, systems, apparatuses, devices, and computer program products are described. An intelligent routing system may route a data object to a path in a process flow using a model, such as a machine-learned model. The system may receive a first data object and may route the first data object along a path of the process flow using a random routing procedure, for example, for model training. The routing may involve performing operations based on the path and the features of the first data object. The system may update one or more models based on an outcome of the operations. Following training, the system may insert a model into the process flow at a decision point between paths. The system may receive a second data object and may route the second data object to a path using the model and based on features of the second data object.