Blockchain Behavior Labeling With Validated Training Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional systems fail to provide comprehensive blockchain operation labeling due to time-consuming and technically complex processes, often relying on datasets with low labeling accuracy, leading to inefficient and compromised machine learning models.
Innovation Solution
A system utilizing a second training dataset comprising independently labeled and validated blockchain operations previously processed through the network, combined with advanced machine learning algorithms, enables real-time, accurate classification and security actions for suspicious blockchain characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional systems use datasets with low labeling accuracy from multiple sources, then the system can process more blockchain operations, but the machine learning model becomes compromised and less reliable
Solution Approach 1:
The patent segments the training data into two distinct datasets: a first training dataset from multiple external sources and a second training dataset from independently labeled and validated blockchain operations. This segmentation allows the system to use diverse data sources while maintaining a core set of high-quality validated data, resolving the contradiction between processing volume and reliability.
Solution Approach 2:
The patent changes the quality parameter of training data by introducing independently labeled and validated blockchain operations as a second training dataset. This parameter change ensures that despite using multiple data sources, the model is trained on high-quality data, maintaining reliability while enabling comprehensive operation processing.
2Measurement precision
If conventional systems manually label blockchain operations, then the labeling can be accurate, but the process becomes time-consuming and complex
Solution Approach 1:
The patent performs preliminary labeling and validation of blockchain operations before they are used for training the machine learning model. By pre-labeling and validating operations in advance, the system ensures high labeling accuracy without requiring time-consuming manual labeling during operation, thus resolving the contradiction between precision and time loss.
Solution Approach 2:
The system uses the blockchain network's own validated operations as training data, allowing the network to self-validate and self-label its operations. This self-service approach eliminates the need for external manual labeling, maintaining accuracy while reducing time and complexity.
3Quantity of substance
If the system uses a single dataset from multiple sources, then data availability increases, but labeling accuracy decreases leading to compromised models
Solution Approach 1:
The patent creates a composite training dataset structure by combining a first training dataset from multiple sources with a second training dataset of independently validated operations. This composite approach allows the system to leverage the quantity and diversity of multiple sources while ensuring the quality and accuracy of the validated subset, resolving the contradiction between data availability and labeling precision.
Data Source
AI summary
Methods and systems use a second dataset comprising independently validated data based on labeled blockchain operations previously processed through the blockchain network. The use of the second dataset in conjunction with the dataset comprising labeled blockchain characteristics data received from a plurality of sources provides a comprehensive input for a machine learning model to identify and label suspicious blockchain operations. By doing so, the system mitigates fraudulent, criminal, or suspicious transactions. For example, the system may send an alert to a user if a behavior type is deemed a high risk by the machine learning model; the system may furthermore hold the funds associated with the blockchain operation in a separate account while the transaction is manually reviewed. Thus, the system may generate alerts to notify a user of a suspicious behavior type or suspicious blockchain operation, divert funds associated with the transaction, and conduct a manual review of suspicious blockchain operations.