Bayesian Multi-Armed Bandit Send Time Model

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional digital communication distribution systems suffer from inaccuracies, inefficiencies, and inflexibility in determining optimal send times for electronic messages due to reliance on heuristic approaches, Monte Carlo simulations, and rigid population-based timing determinations, which lead to suboptimal performance and resource wastage.

Innovation Solution

The implementation of a predictive communications distribution system using Bayesian approaches and multi-armed bandit algorithms to dynamically balance data from different user attribute groups and time granularities, employing a Bayes upper-confidence-bound send time model to balance exploration and exploitation, and applying constraint regression for accurate individualized send time determination.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If conventional systems utilize heuristic approaches and pre-determined rules to determine distribution times, then the systems are easy to implement, but the accuracy of timing determination deteriorates

Engineering Contradiction:

Improveease of implementationVSAvoidaccuracy of timing determination

Core Design Contradiction:

Ease of manufactureVSMeasurement precision

Solution Approach 1:

The patent replaces conventional heuristic and rule-based mechanical systems with machine learning models that automatically learn optimal distribution timing patterns from historical data, substituting manual rule formulation with automated data-driven decision-making algorithms

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system dynamically adjusts distribution timing parameters based on learned patterns from historical data, transforming static pre-determined rules into adaptive parameters that change based on observed user behavior and response patterns

Inventive Principle:

Principle #35Parameter changes

2Reliability

If conventional systems utilize Monte Carlo simulations to determine timing, then comprehensive analysis is achieved, but computational expense and time increase significantly

Engineering Contradiction:

Improvecomprehensiveness of analysisVSAvoidcomputation time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system performs preliminary learning during offline training phases where machine learning models are trained on historical data to capture timing patterns, so that during online operation only lightweight inference is needed rather than repeated comprehensive simulations

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent creates simplified surrogate models through machine learning that replicate the behavior of complex Monte Carlo simulations, allowing fast approximation of timing decisions without performing computationally expensive simulations in real-time

Inventive Principle:

Principle #26Copying

3Stability of the object's composition

If conventional systems apply rigid population-based timing determinations, then consistency across users is maintained, but adaptability to individual user behavior deteriorates

Engineering Contradiction:

Improveconsistency of timing approachVSAvoidadaptability to individual behavior

Core Design Contradiction:

Stability of the object's compositionVSAdaptability or versatility

Solution Approach 1:

The patent segments the user population into different groups or individuals, applying personalized timing strategies to each segment while maintaining overall system consistency through a unified machine learning framework that handles multiple user profiles

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system applies different timing characteristics to different user segments or individuals, allowing local optimization for each user group while maintaining global system coherence through the unified learning model

Inventive Principle:

Principle #3Local quality

4Quantity of substance

If conventional systems rely on broad-spectrum population-based information, then data availability is improved, but individualized timing accuracy deteriorates

Engineering Contradiction:

Improveavailability of dataVSAvoidindividualized timing accuracy

Core Design Contradiction:

Quantity of substanceVSMeasurement precision

Solution Approach 1:

The patent merges population-level data with individual user data in a unified machine learning model, combining the advantages of abundant population data with the need for personalized timing accuracy through hierarchical or multi-level modeling approaches

Inventive Principle:

Principle #5Merging (Combining)

Data Source

PatentUS11710065B2Utilizing a bayesian approach and multi-armed bandit algorithms to improve distribution timing of electronic communications

Publication Date: 2023.07.25 ADOBE INC

AI summary

The present disclosure relates to systems, methods, and non-transitory computer readable media for determining send times to provide electronic communications based on predicted response rates by utilizing a Bayesian approach and multi-armed bandit algorithms. For example, the disclosed systems can generate predicted response rates by training and utilizing one or more response rate prediction models to generate a weighted combination of user-specific response information and population-specific response information. The disclosed systems can further utilize a Bayes upper-confidence-bound send time model to determine send times that are more likely to elicit user responses based on the predicted response rates and further based on exploration and exploitation considerations. In addition, the disclosed systems can update the response rate prediction models and/or the Bayes upper-confidence-bound send time model based on providing additional electronic communications and receiving additional responses to modify model weights.