ML-Based Communication Scheme Selection for MFA
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In multi-factor authentication systems, inactive user devices due to unpaid bills or missed payments lead to wasted computing resources as MFA servers send messages that are not received, causing incorrect fraud detection and additional authentication attempts, which affects resource utilization and accuracy.
Innovation Solution
A prediction and monitoring system using a machine learning model identifies recurring events and predicts transaction dates and amounts, determining the device's status to select appropriate communication schemes, ensuring messages are successfully delivered and reducing unnecessary resource usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If MFA servers send messages to user devices for authentication, then authentication process is initiated, but computing resources are wasted when devices are inactive due to unpaid bills
Solution Approach 1:
The system performs preliminary actions by predicting future transaction dates and determining device status before authentication is needed. The machine learning model analyzes historical payment behavior to forecast when transactions are likely to occur and proactively determines whether a device is active or inactive, allowing the system to prepare appropriate response strategies in advance without wasting resources on definitely inactive devices.
Solution Approach 2:
The system implements feedback loops by continuously monitoring actual transaction behavior against machine learning predictions. The feedback mechanism compares predicted transaction dates with actual transactions, refining the machine learning model's accuracy over time. This feedback enables more precise device status determination and communication strategy selection, reducing resource waste while maintaining authentication reliability.
2Productivity
If messages are sent to inactive user devices, then authentication attempts are made, but fraud detection accuracy is compromised due to missed messages
Solution Approach 1:
The system performs preliminary device status determination using machine learning predictions before initiating authentication sequences. By predicting transaction dates and analyzing payment behavior patterns in advance, the system can accurately identify inactive devices before sending messages, preventing wasted authentication attempts on devices that will not receive or process messages, thereby maintaining fraud detection accuracy.
Solution Approach 2:
The system uses feedback from actual message delivery results to refine device status determination. By monitoring whether messages are successfully delivered and processed, the system adjusts its predictions and device status classifications accordingly. This feedback mechanism continuously improves fraud detection accuracy by learning from actual user behavior patterns and message delivery outcomes.
3Reliability
If multiple authentication attempts are made, then user access is ensured, but additional computing resources are consumed
Solution Approach 1:
The system performs preliminary device status determination and communication channel validation before initiating authentication sequences. By predicting which devices are likely to be active and capable of receiving messages, the system can pre-filter authentication attempts to only those likely to succeed, reducing unnecessary repetition and complexity while ensuring reliable user access through targeted authentication efforts.
Solution Approach 2:
The system dynamically adjusts authentication strategies based on real-time device status determination and machine learning predictions. Rather than using static authentication protocols, the system adapts its approach by selecting appropriate communication channels and authentication methods based on the predicted and actual device state, optimizing the balance between access assurance and process complexity.
Data Source
AI summary
In some implementations, a prediction and monitoring system may processing, using a machine learning model, account data associated with an account that is associated with a user of a user device to identify a series of recurring events associated with the user device. The prediction and monitoring system may generate, using the machine learning model, a predicted transaction date and a predicted transaction amount that are both associated with the series of recurring events. The prediction and monitoring system may select, based on additional account data associated with the account and at least one of the predicted transaction date or the predicted transaction amount, a particular communication scheme, of a plurality of communication schemes, for communicating with the user. The prediction and monitoring system may transmit at least one message according to the particular communication scheme to facilitate authentication of the user.