Scalable Multimodal HTS Classification Through Probability Aggregation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Misclassification of Harmonized Tariff Schedule (HTS) codes leads to customs penalties, border delays, and product seizures, highlighting the need for accurate and efficient code classification systems.
Innovation Solution
A system utilizing trained text and image classification models to determine code designations, combining textual and image data to generate candidate code designations, and aggregating probabilities to ensure high-confidence selections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional manual classification methods are used for HTS codes, then flexibility and adaptability to complex items are maintained, but classification accuracy and efficiency deteriorate leading to misclassification penalties and border delays
Solution Approach 1:
The patent segments the code classification task into multiple independent classification models, each specialized for specific product categories or code ranges. This segmentation allows parallel processing of different item types, improving overall efficiency while maintaining high accuracy through specialized models trained on category-specific data patterns
Solution Approach 2:
The patent replaces manual mechanical classification processes with automated machine learning classification systems. These systems use trained models to automatically analyze item descriptions, specifications, and attributes, substituting human judgment with algorithmic decision-making that operates at higher speeds and scales to handle large volumes of classifications without fatigue or inconsistency
2Productivity
If automated classification systems are implemented to improve efficiency, then processing speed increases, but classification accuracy may deteriorate due to inability to handle complex or novel items
Solution Approach 1:
The patent implements dynamic classification systems where models can be retrained and updated with new data and emerging product types. The system adapts to novel items by incorporating feedback from manual reviews and continuously learning from new classification cases, allowing it to maintain high accuracy while processing efficiency remains high through automated operations
Solution Approach 2:
The patent incorporates feedback mechanisms where classification results are reviewed and corrected when necessary, with these corrections fed back into the training data to improve future automated classifications. This feedback loop ensures that the system learns from its mistakes and improves accuracy over time while maintaining high processing throughput
3Measurement precision
If multiple classification models are used to improve accuracy, then classification precision increases, but system complexity and computational resources required deteriorate
Solution Approach 1:
The patent segments the classification system into modular, independent models that can be selectively applied based on item characteristics. This segmentation reduces overall system complexity by allowing only relevant models to be activated for each classification task, rather than running all models on every item, thereby managing computational resources efficiently while maintaining high accuracy through specialized models
Data Source
AI summary
Systems and methods of determining and transmitting a code designation are disclosed. A request for a code designation of an item, and information about the item are received. A first probability distribution is obtained via a trained text classification model from the textual description. A second probability distribution is obtained via a trained image classification model, from the one or more images. A plurality of candidate code designations based on the first probability distribution and the second probability distribution are generated. Each candidate code designation includes a first portion and a second portion of the code designation. Respective probabilities of respective candidate code designations in a first subset of the plurality of candidate code designations are aggregated, from which a selected code designation is transmitted to the requesting system as the code designation associated with the item.