Online Machine Learning with Drift Detection 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 continuously and be understandable by domain experts.
Innovation Solution
A continuously learning, stable, and robust online machine learning system (OMLS) that learns concurrently from both streaming data and human-provided feedback, incorporating an Online Machine Learning Engine (OMLE), Online Preprocessing Engine (OPrE), Online Feature Engineering Engine (OFEE), and Online Robust Feature Selection Engine (ORFSE), along with an Online Explanation System (OES) and Online Human Expert Feedback System (OEFS) to provide transparent and updatable models.
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 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, transforming the static batch training approach into a dynamic adaptive system. The model evolves over time to track changing data distributions while maintaining performance through incremental parameter adjustments.
Solution Approach 2:
The system incorporates feedback mechanisms where model predictions are continuously evaluated against actual outcomes, and this feedback is used to update model parameters in real-time. This closed-loop feedback enables the model to self-correct and adapt to changing patterns without manual retraining.
2Adaptability or versatility
If models are periodically retrained to adapt to changing data, then adaptability improves, but model stability decreases and business continuity is disrupted
Solution Approach 1:
The patent implements continuous online learning where the model learns indefinitely from incoming data streams without interruption. This eliminates the periodic retraining cycles that cause instability, maintaining continuous model evolution while ensuring uninterrupted business operations and consistent model behavior.
3Measurement precision
If complex machine learning models are used to capture emerging patterns, then prediction accuracy improves, but model understandability and interpretability decrease
Solution Approach 1:
The patent segments the complex model into interpretable components, including feature importance analysis, decision rules extraction, and component-wise explanations. This allows the system to maintain high predictive accuracy through complex interactions while providing transparent explanations of individual feature contributions and decision logic.
4Adaptability or versatility
If online learning is implemented to learn continuously from streaming data, then adaptability to changing patterns improves, but model stability and robustness may deteriorate
Solution Approach 1:
The patent incorporates regularization techniques and constraints that act as preemptive measures against overfitting and instability. These cushioning mechanisms are built into the online learning process to prevent the model from becoming too sensitive to noise or adapting too aggressively to changing patterns, thereby maintaining stability while preserving adaptability.
Data Source
AI summary
An Online Machine Learning System (OMLS) including an Online Preprocessing Engine (OPrE) configured to (a) receive streaming data including an instance comprising a vector of inputs, the vector of inputs comprising a plurality of continuous or categorical features; (b) discretize features; (c) impute missing feature values; (d) normalize features; and (e) detect drift or change in features; an Online Feature Engineering Engine (OFEE) configured to produce features; and an Online Robust Feature Selection Engine (ORFSE) configured to evaluate and select features; an Online Machine Learning Engine (OMLE) configured to incorporate and utilize 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 each of the OMLE, the OPrE, the OFEE, and the ORFSE are continuously communicatively coupled to each other, and wherein the OMLS is configured to perform continuous online machine learning.