Atypical Replacement Explanations for Out-of-Stock Recommendations

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current online recommendation systems often suggest popular items that are likely to be out of stock, failing to provide atypical replacement options that users may not be aware of due to their knowledge gaps, leading to decision fatigue and an unsatisfactory shopping experience.

Innovation Solution

An online system uses a machine-learned language model to generate explanations for atypical replacement items, recommending alternatives that serve the same purpose as the target item and providing contextualized reasons for their selection, thereby guiding users to suitable substitutes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If the recommendation system recommends popular items based on previous purchase history, then the recommendation accuracy is improved, but the availability of recommended items deteriorates because they are more likely to be out of stock

Engineering Contradiction:

Improverecommendation accuracyVSAvoiditem availability

Core Design Contradiction:

Measurement precisionVSReliability

Solution Approach 1:

Instead of recommending only popular items (high probability of selection), the system inverts the approach by identifying and recommending atypical replacement items (low probability of selection) that users may not be aware of. This is achieved by using a language model to generate explanations for why certain atypical items could serve as replacements, thereby converting the harmful effect of popularity-based recommendations into a benefit by highlighting lesser-known alternatives.

Inventive Principle:

Principle #13The other way round (Inversion)

Solution Approach 2:

The patent introduces a language model as an intermediary between the recommendation system and the user. This intermediary generates contextualized explanations for atypical replacement items, bridging the knowledge gap between the system's data-driven insights and the user's understanding. The language model translates complex recommendation logic into human-readable explanations that guide users toward suitable substitutes.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Reliability

If the system recommends atypical replacement items, then the item availability is improved, but the user awareness of these alternatives deteriorates due to knowledge gaps

Engineering Contradiction:

Improveitem availabilityVSAvoiduser knowledge

Core Design Contradiction:

ReliabilityVSLoss of information

Solution Approach 1:

The language model serves as an intermediary that compensates for user knowledge gaps by generating contextualized explanations for atypical replacement items. It bridges the information asymmetry between the system's awareness of alternatives and the user's knowledge, translating technical recommendation logic into understandable rationale that guides user decision-making.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system changes the parameter of information presentation by transforming raw recommendation data into natural language explanations. Instead of simply listing atypical alternatives, the system modifies the information parameter to include contextualized reasoning, thereby making previously invisible alternatives comprehensible and actionable for users.

Inventive Principle:

Principle #35Parameter changes

3Loss of time

If the system provides detailed explanations for replacement items, then the user decision-making speed is improved, but the system complexity increases due to language model integration

Engineering Contradiction:

Improveuser decision timeVSAvoidsystem architecture

Core Design Contradiction:

Loss of timeVSDevice complexity

Solution Approach 1:

The language model acts as an intermediary layer that handles the complex task of generating explanations, isolating the complexity from the core recommendation logic. This mediator translates structured recommendation data into natural language explanations without requiring changes to the underlying recommendation algorithm, thereby managing system complexity while enhancing user experience.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system segments the recommendation process into distinct modules: the core recommendation engine that identifies atypical alternatives, and the language model that generates explanations. This segmentation allows each component to specialize in its function, reducing overall system complexity by creating modular, independently manageable units with clear interfaces.

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS12536184B2Generating explanations for atypical replacements using large language machine-learned models

Publication Date: 2026.01.27 MAPLEBEAR INC

AI summary

An online system performs an atypical replacement recommendation task in conjunction with a model serving system or the interface system to make recommendations to a user for replacing a target item with an atypical replacement item. The online system receives a search query from a user and identifies a target item based on the search query. The online system identifies a set of candidate items for replacing the target item. The online system may select one or more atypical replacement items in the set of candidate items, and generate an explanation for each atypical replacement item. The explanation provides a reason for using the atypical replacement item to replace the target item. The online system provides the atypical replacement items and the corresponding explanations as a response to the search query.