Online Machine Learning With Explanations and Expert Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current machine learning systems face challenges in maintaining stability and robustness in dynamic environments, such as finance, telecommunications, and health, where data continuously changes, and there is a need for models that can learn from both data and human feedback in a user-friendly manner, while also being understandable by domain experts.
Innovation Solution
A continuously learning, stable, and robust online machine learning system (OMLS) that incorporates an Online Machine Learning Engine (OMLE) with Online Preprocessing, Feature Engineering, and Robust Feature Selection Engines, allowing concurrent learning from streaming data and human input, and an Online Explanation System for model transparency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If machine learning models are trained in batch mode on available datasets, then model performance can be optimized on training data, but the models fail to adapt to continuously changing real-world data patterns and require periodic retraining
Solution Approach 1:
The patent implements dynamic online learning where model parameters are continuously updated as new data arrives, transitioning from static batch training to dynamic adaptive learning. The system processes data streams in real-time, allowing the model to adapt its structure and parameters dynamically to changing patterns while maintaining performance stability through controlled learning rates and validation mechanisms.
Solution Approach 2:
The system incorporates feedback loops where model predictions are continuously evaluated against actual outcomes, and this feedback is used to update model parameters in real-time. The feedback mechanism includes performance monitoring, error tracking, and automated retraining triggers that ensure the model adapts to changing data distributions while maintaining reliability through validation checks.
2Measurement precision
If complex machine learning models are used to capture emerging patterns in data, then prediction accuracy improves, but the models become difficult to understand and operate for business people
Solution Approach 1:
The patent introduces explanation systems and visualization interfaces that act as intermediaries between complex model internals and business users. These intermediaries translate complex model decisions into understandable explanations, feature importance visualizations, and business-relevant metrics, allowing users to understand and operate with complex models without needing to comprehend the underlying mathematical complexity.
3Reliability
If models are periodically retrained to maintain performance, then model accuracy can be preserved, but business continuity is disrupted and it is difficult to know when retraining should occur
Solution Approach 1:
The system implements continuous online learning that operates without interruption to business processes. Data is processed and model parameters are updated continuously as new information arrives, eliminating the need for periodic batch retraining that disrupts operations. The useful action of learning and adaptation continues uninterrupted, maintaining both accuracy and business continuity simultaneously.
Solution Approach 2:
The system incorporates automated monitoring and self-triggered retraining mechanisms that detect performance degradation or data distribution changes and initiate model updates automatically without human intervention. This self-service capability ensures model accuracy is maintained while eliminating the need for manual retraining scheduling, thereby preserving business continuity.
4Adaptability or versatility
If online learning models continuously adapt to changing datasets, then they can handle dynamic environments, but it becomes difficult to track model decisions and maintain accountability
Solution Approach 1:
The system maintains copies of model states, parameter versions, and decision histories at each update point. These copies are stored in version-controlled repositories that allow reconstruction of any past model state and its decisions. This copying mechanism enables full trackability of model decisions over time while the model continues to adapt continuously to changing data, preserving accountability information alongside adaptation capability.
Data Source
AI summary
An Online Machine Learning System (OMLS) including an Online Explanation System (OES), updated continuously, configured to provide instance level explanations and model level explanations to a user; an Online Human Expert Feedback System (OEFS), updated continuously, configured to obtain expert instance level feedback and expert model level feedback, for optimization of operation of the OMLS; an Online Machine Learning Engine (OMLE) for incorporating and utilizing one or more machine learning algorithms or models utilizing features to generate a result, and capable of incorporating and utilizing multiple different machine learning algorithms or models, wherein the OMLS is configured to perform continuous online machine learning.