Two-Phase HS and HTS Code Classification System
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing classification tools for Harmonized System (HS) and Harmonized Tariff Schedule (HTS) codes face challenges in balancing global accuracy with country-specific HTS code coverage, as they struggle to share information across markets while maintaining nuanced market-specific classifications, leading to high costs and inefficiencies for merchants in cross-border commerce.
Innovation Solution
A two-phase approach is implemented, where a global classification model using a bespoke static set of classes generates HS codes, followed by a locally specific zero- or few-shot classification model for HTS refinement, allowing for scalable and accurate country-specific HTS code generation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single-phase zero-shot classification model is used for HTS code generation, then country-specific classification accuracy is improved, but processing speed and computational efficiency deteriorate
Solution Approach 1:
The classification process is segmented into two distinct phases: a first phase using a global classification model for HS code generation, and a second phase using a locally specific model for HTS code refinement. This segmentation allows each phase to be optimized independently, with the global model handling broad categorization efficiently and the local model providing precise country-specific classification only when needed.
Solution Approach 2:
The system performs preliminary classification at the HS code level using a globally trained model before proceeding to the more time-consuming HTS code refinement phase. This preliminary action establishes a solid foundation for country-specific classification while avoiding the computational burden of performing full zero-shot classification from scratch for every country-specific code.
2Productivity
If a global classification model with static classes is used, then computational efficiency is improved, but adaptability to country-specific nuances deteriorates
Solution Approach 1:
The system employs different classification models with different characteristics for different stages of the classification process. The global model uses a static set of classes optimized for computational efficiency and broad applicability, while the local model uses a dynamic set of classes adapted to country-specific tariff schedules and classification nuances, ensuring each model operates with optimal quality for its specific function.
Solution Approach 2:
The classification system transitions from a static classification framework at the global HS code level to a dynamic framework at the local HTS code level. The global model with its fixed class structure provides stable, efficient processing, while the local model adapts its classification categories based on specific country requirements, enabling the system to handle both efficiency and adaptability requirements at appropriate levels.
3Productivity
If a two-phase classification approach is implemented, then classification coherence and speed are improved, but system complexity increases
Solution Approach 1:
The classification system is divided into two sequential phases with distinct models: the first phase uses a global classification model for HS code generation, and the second phase uses a locally specific model for HTS code refinement. This segmentation allows the system to achieve improved speed and coherence by processing classification tasks in staged manner, with each phase optimized for its specific objectives.
Solution Approach 2:
The HS code classification result serves as an intermediary output from the first phase that feeds into the second phase. This intermediary mechanism allows the system to leverage the efficient global classification results as a foundation for the more precise local classification, reducing the overall complexity burden by building upon rather than replacing the first phase's work.
Data Source
AI summary
Systems for providing globally optimal and locally specific classification of a plurality of goods using Harmonized System (HS) codes and Harmonized Tariff Schedule (HTS) codes are disclosed. Some embodiments include: maintaining a database of import/export classifications for a plurality of goods comprising Harmonized System codes of the plurality of goods and Harmonized Tariff Schedule codes of the plurality of goods, receiving an import/export classification request from a user for an import/export classification of a good of a cross-border commerce transaction, responding to the import/export classification request by generating a computerized classification result by searching the database of import/export classifications for the plurality of goods using a two phase approach and providing a Harmonized Tariff Schedule code for the good of the cross-border commerce transaction to the user in response to the import/export classification request with increased Coherence and speed.