DATE: Dual Attentive Tree-aware Embedding for Customs Fraud Detection

2020-08-23 · KDD 2020 8 · Sundong Kim, Yu-Che Tsai, Karandeep Singh, Yeonsoo Choi, Etim Ibok, Cheng-Te Li, Meeyoung Cha

Intentional manipulation of invoices that lead to undervaluation of trade goods is the most common type of customs fraud to avoid ad valorem duties and taxes. To secure government revenue without interrupting legitimate trade flows, customs administrations around the world strive to develop ways to detect illicit trades. This paper proposes DATE, a model of Dual-task Attentive Tree-aware Embedding, to classify and rank illegal trade flows that contribute the most to the overall customs revenue when caught. The strength of DATE comes from combining a tree-based model for interpretability and transaction-level embeddings with dual attention mechanisms. To accurately identify illicit transactions and predict tax revenue, DATE learns simultaneously from illicitness and surtax of each transaction. With a five-year amount of customs import data with a test illicit ratio of 2.24%, DATE shows a remarkable precision of 92.7% on illegal cases and a recall of 49.3% on revenue after inspecting only 1% of all trade flows. We also discuss issues on deploying DATE in Nigeria Customs Service, in collaboration with the World Customs Organization.

📄 PDF Abstract BibTeX

Code (1)

Roytsai27/Dual-Attentive-Tree-aware-Embedding pytorch

Tasks

Fraud Detection Multi-target regression Value prediction

Methods 이 논문이 사용한 방법론

Attention 설명 없음

Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…

Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.

Interpretability 설명 없음

Multi-Head Attention 설명 없음

Learning to Embed Sentences Using Attentive Recursive Trees

2018-11-06 · Jiaxin Shi, Lei Hou, Juanzi Li, Zhiyuan Liu 외

Sentence embedding is an effective feature representation for most deep learning-based NLP tasks. One prevailing line of methods is using recursive latent tree-structured networks to embed sentences with task-specific st…

Sentence Sentence Embedding Sentence-Embedding

Dependency-aware Prototype Learning for Few-shot Relation Classification

2022-10-01 · COLING 2022 10 · Tianshu Yu, Min Yang, Xiaoyan Zhao

Few-shot relation classification aims to classify the relation type between two given entities in a sentence by training with a few labeled instances for each relation. However, most of existing models fail to distinguis…

Classification Few-Shot Relation Classification Relation Relation Classification+1

Generalized Stock Price Prediction for Multiple Stocks Combined with News Fusion

2026-03-08 · Pei-Jun Liao, Hung-Shin Lee, Yao-Fei Cheng, Li-Wei Chen 외 arxiv

Predicting stock prices presents challenges in financial forecasting. While traditional approaches such as ARIMA and RNNs are prevalent, recent developments in Large Language Models (LLMs) offer alternative methodologies…

Stock Price Prediction

Phoneme-aware and Channel-wise Attentive Learning for Text DependentSpeaker Verification

2021-06-25 · Yan Liu, Zheng Li, Lin Li, Qingyang Hong

This paper proposes a multi-task learning network with phoneme-aware and channel-wise attentive learning strategies for text-dependent Speaker Verification (SV). In the proposed structure, the frame-level multi-task lear…

Multi-Task Learning Speaker Verification Text-Dependent Speaker Verification

Urban Region Representation Learning with Attentive Fusion

2023-12-07 · Fengze Sun, Jianzhong Qi, Yanchuan Chang, Xiaoliang Fan 외

An increasing number of related urban data sources have brought forth novel opportunities for learning urban region representations, i.e., embeddings. The embeddings describe latent features of urban regions and enable d…

Representation Learning