Explainable Product Classification Model
The task of assigning internationally accepted commodity codes (aka HS code) to traded goods is a critical function of customs offices. Like court decisions made by judges, this task follows the doctrine of precedent and can be nontrivial even for experienced officers. Together with the Korea Customs Service (KCS), we propose a first-ever explainable decision supporting model that suggests the most likely subheadings (i.e., the first six digits) of the HS code. The model also provides reasoning for its suggestion in the form of a document that is interpretable by customs officers. We evaluated the model using 5,000 cases that recently received a classification request. The results showed that the top-3 suggestions made by our model had an accuracy of 93.9% when classifying 925 challenging subheadings. A user study with 32 customs experts further confirmed that our algorithmic suggestions accompanied by explainable reasonings can substantially reduce the time and effort taken by customs officers for classification reviews.
Try HS Classification with your data:
https://bit.ly/IBS-HS-Classification
Feedback from field officers:
“The model gave suggestions that I could have missed. I found this very helpful.”
“The supporting model gave a rough idea of final decision I had to make.”
“The model helped me make quick decisions.”
“Since the model shows the candidates, it can be helpful to educate new workers who have short working experience and expertise in the classification task.”
For a glance at our current work, have a look at our paper:
[1] Classification of Goods Using Text Descriptions With Sentences Retrieval (KAIA 2020): [Link]
[2] Explainable Product Classification for Customs (ACM TIST): [Link]
If you are interested in this project and interpretable AI, feel free to contact us:
Sundong Kim (sundong@ibs.re.kr): Project lead
Eunji Lee (mk35471@kaist.ac.kr): Algorithm and modeling
Sihyeon Kim (sihk@kaist.ac.kr): Experimentation and web service