User-Treatment Cost Prediction for Closing Interaction Gaps

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Online systems face challenges in selecting users for treatments due to varying user responses and dynamic environmental variables, leading to insufficient user interactions, particularly when predicting interaction rates that change over time.

Innovation Solution

An online system computes incremental cost predictions for user-treatment pairs using an interaction model and a cost model to select users and treatments effectively, addressing interaction gaps and reducing costs.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If treatments are applied to all users to ensure sufficient interactions, then user interaction quantity is improved, but system cost increases

Engineering Contradiction:

Improveuser interaction quantityVSAvoidsystem cost

Core Design Contradiction:

ProductivityVSLoss of energy

Solution Approach 1:

The patent applies local quality by differentiating treatment application based on individual user characteristics and predicted response probabilities. Instead of uniform treatment distribution, the system identifies specific user segments with higher likelihood to respond positively to treatments, applying interventions selectively to those users. This resolves the contradiction by maintaining sufficient interaction quantity through targeted treatment of high-response users while reducing overall system cost by avoiding treatments for low-response users.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The patent utilizes parameter changes by dynamically adjusting treatment selection based on predicted interaction probabilities and cost parameters. The system continuously updates user profiles and treatment effectiveness parameters, then reoptimizes treatment allocation based on current parameter states. This allows the system to adapt treatment application to changing user behaviors and environmental conditions, maintaining interaction quality while optimizing cost efficiency.

Inventive Principle:

Principle #35Parameter changes

2Loss of energy

If treatments are applied to users with high predicted interaction probability, then cost efficiency is improved, but user interaction quantity may be insufficient

Engineering Contradiction:

Improvecost efficiencyVSAvoiduser interaction quantity

Core Design Contradiction:

Loss of energyVSProductivity

Solution Approach 1:

The patent applies preliminary action by pre-calculating and storing user treatment response probabilities and cost parameters before actual treatment deployment. The system performs preliminary modeling and segmentation of users based on historical data, then uses these pre-computed parameters to guide real-time treatment selection. This allows the system to efficiently balance cost efficiency and interaction quantity by relying on pre-established user profiles rather than evaluating each user from scratch during treatment allocation.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent implements dynamics by making treatment allocation adaptive and responsive to changing conditions. The system continuously monitors user responses, environmental variables, and treatment effectiveness, then dynamically adjusts treatment selection parameters. This dynamic approach allows the system to respond to emerging patterns in user behavior and external factors, ensuring optimal balance between cost efficiency and interaction quantity under varying conditions.

Inventive Principle:

Principle #15Dynamics

3Ease of manufacture

If the system selects users based on static treatment models, then implementation simplicity is improved, but adaptability to changing conditions deteriorates

Engineering Contradiction:

Improveimplementation simplicityVSAvoidadaptability to changing conditions

Core Design Contradiction:

Ease of manufactureVSAdaptability or versatility

Solution Approach 1:

The patent applies feedback by implementing continuous monitoring and evaluation of treatment effectiveness, user response patterns, and environmental changes. The system uses this feedback information to iteratively refine user profiles, update treatment selection models, and adjust allocation strategies. This feedback mechanism maintains implementation simplicity by automating the refinement process through machine learning models that continuously learn from operational data, while significantly improving adaptability to changing conditions through data-driven model updates.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS20260080424A1Incremental cost prediction for user treatment selection

Publication Date: 2026.03.19 MAPLEBEAR INC

AI summary

An online system computes an incremental cost prediction for each of a set of user-treatment pairs to select a set of treatments to apply to users to satisfy a predicted interaction gap. The online system generates a set of candidate user-treatment pairs that each include user data for a user of the online system and treatment data for a treatment of a set of treatments. The online system computes an incremental interaction prediction and a treatment cost prediction for each of the candidate user-treatment pairs by applying an incremental interaction model to the user data and the treatment data in each user-treatment pair. The online system computes incremental cost predictions for each of the user-treatment pairs based on the computed incremental interaction predictions and treatment cost predictions and selects which users to apply treatments to and which treatments to apply to those users based on the incremental cost predictions.