Item Identifier Voting for Low-Resource Smart Cart Prediction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Smart systems, such as smart shopping carts, face limitations in computational resources, making it challenging to effectively utilize machine-learning models for item identification due to memory and processing demands.

Innovation Solution

Implementing an efficient selection algorithm that minimizes computational resources by selecting item identifier predictions from multiple machine-learning models independently, using algorithms like simple majority or weighted voting.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If multiple machine-learning models are used to generate item identifier predictions, then prediction accuracy is improved, but computational resource consumption increases

Engineering Contradiction:

Improveprediction accuracyVSAvoidcomputational resource consumption

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent segments the prediction task into multiple independent machine-learning models, each generating predictions separately. This allows the system to leverage diverse model strengths for improved accuracy while keeping each model computationally manageable. The segmentation of prediction generation from selection (via voting algorithm) resolves the contradiction by distributing computational load across independent components.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces a voting algorithm as an intermediary mechanism between multiple machine-learning models and the final item identifier selection. This intermediary aggregates predictions from multiple models without requiring a single complex model, achieving improved accuracy through ensemble methods while maintaining computational efficiency through simple voting logic rather than complex integrated processing.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Measurement precision

If a machine-learning model is trained to select item identifier predictions from multiple models, then selection accuracy is improved, but memory requirements increase

Engineering Contradiction:

Improveselection accuracyVSAvoidmemory requirements

Core Design Contradiction:

Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent replaces a complex machine-learning selection model with simple voting algorithms (majority voting or weighted voting) that require minimal memory. These lightweight selection mechanisms achieve sufficient accuracy without the heavy memory footprint of trained models, effectively using computationally inexpensive methods to resolve predictions.

Inventive Principle:

Principle #27Cheap short-living objects (Disposable)

Solution Approach 2:

The patent substitutes a machine-learning-based selection system with a deterministic voting algorithm system. This replacement eliminates the need for training and storing selection model parameters, significantly reducing memory requirements while maintaining selection functionality through algorithmic aggregation of model predictions.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Productivity

If computational resources are allocated to the selection algorithm, then selection performance is improved, but resources available for machine-learning models decrease

Engineering Contradiction:

Improveselection performanceVSAvoidresources available for machine-learning models

Core Design Contradiction:

ProductivityVSPower

Solution Approach 1:

The patent applies partial action by using simple voting algorithms that require minimal computational resources for selection, leaving sufficient resources for the machine-learning models. Rather than investing heavily in a complex selection model, the system uses just enough selection logic (voting) to aggregate predictions effectively, preserving resources for the more resource-intensive prediction generation phase.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS20260017969A1Prediction selection for item identifiers using efficient selection algorithm

Publication Date: 2026.01.15 MAPLEBEAR INC

AI summary

A smart system, such as a smart shopping cart system, uses an efficient selection algorithm to select an item identifier prediction for an item. The smart cart system uses a set of machine-learning models to generate identifier predictions based on images. To select an item identifier, the smart system applies an efficient selection algorithm to the predictions from the machine-learning models. An efficient selection algorithm is an algorithm that requires minimal computational resources to perform. For example, the efficient selection algorithm may be a simple majority algorithm that selects the identifier prediction generated by a majority of the models or a weighted voting algorithm where each model's vote is weighted by some metric. The smart system applies the efficient selection algorithm to select an item identifier prediction from the ones generated by the models. The smart system may display content related to the item associated with the item identifier prediction.