Yu-Che Tsai

Yu-Che Tsai

Auto-generated profile, based on publicly available information

Yu-Che Tsai

National Taiwan University, Taipei, Taiwan

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
2 September 2025 in ACM Computing Surveys
Verified
Journal Article
Open access
Cheng-Te LiCheng-Te Li
Cheng-Te Li
YTYu-Che Tsai
Yu-Che Tsai
CCChih-Yao Chen
Chih-Yao Chen
JLJay Chiehen Liao
Jay Chiehen Liao
This survey dives into Tabular Data Learning (TDL) using Graph Neural Networks (GNNs), a domain where deep learning-based approaches have increasingly shown superior performance in both classification and regression tasks compared to traditional methods. We highlight a critical gap in deep neural TDL: the underrepresentation of latent correlations among data instances and feature values. GNNs, with their innate capability to model intricate relationships and interactions between diverse elements of tabular data, have garnered significant interest and application across TDL domains. Our survey provides a systematic review of the methods involved in designing and implementing GNNs for TDL (GNN4TDL). It encompasses a detailed investigation into the foundational aspects and an overview of GNN-based TDL methods, offering insights into their evolving landscape. We present a comprehensive taxonomy focused on constructing graph structures and representation learning within GNN-based TDL methods. We also examine various training plans, emphasize the integration of auxiliary tasks to enhance the representation quality. A critical part of our discussion is dedicated to the practical applications across a spectrum of GNN4TDL scenarios, exhibiting their versatility and impact. Last, we discuss the limitations and future directions, aiming at spurring advancements in GNN4TDL. This survey serves as a resource for researchers and practitioners, offering a thorough understanding of GNNs’ role in revolutionizing TDL and pointing toward future innovations in this promising area.

30 April 2023
Verified
Conference Paper
Open access
VFValeria Fionda
Valeria Fionda
OHOlaf Hartig
Olaf Hartig
RAReyhaneh Abdolazimi
Reyhaneh Abdolazimi
SASihem Amer-Yahia
Sihem Amer-Yahia
HCHongzhi Chen
Hongzhi Chen
WZWenwu Zhu
Wenwu Zhu
This paper summarizes the content of the 28 tutorials that have been given at The Web Conference 2023.

21 October 2023
Verified
Conference Paper
Open access
KSKarandeep Singh
Karandeep Singh
YTYu-Che Tsai
Yu-Che Tsai
Cheng-Te LiCheng-Te Li
Cheng-Te Li
MCMeeyoung Cha
Meeyoung Cha
SLShou-De Lin
Shou-De Lin
Customs officials across the world encounter huge volumes of transactions. Associated with customs transactions is customs fraud-the intentional manipulation of goods declarations to avoid taxes and duties. Due to limited manpower, the customs offices can only manually inspect a small number of declarations, necessitating the automation of customs fraud detection by machine learning techniques. The limited availability of manually inspected ground truth data makes it essential for the ML approach to generalize well on unseen data. However, current customs fraud detection models are not well suited or designed for this setting. In this work, we propose GraphFC (Graph Neural networks for Customs Fraud), a model-agnostic, domain-specific, graph neural network based customs fraud detection model that is designed to work in a real-world setting with limited ground truth data. Extensive experimentation using real customs data from two countries demonstrates that GraphFC generalizes well over unseen data and outperforms various baselines and other models by a large margin.

13 May 2024
Verified
Conference Paper
Open access
YLYen-Wen Lu
Yen-Wen Lu
YTYu-Che Tsai
Yu-Che Tsai
Cheng-Te LiCheng-Te Li
Cheng-Te Li
As the digital commerce landscape continues to expand with rating platforms, the consumer base has similarly grown, marking a pivotal reliance on user ratings and reviews. However, the rise of fraudulent users leveraging deceitful conduct, such as manipulation of product rankings, challenges the credibility of these platforms and compels the necessity for an effective detection model. Amid the challenges of evolving fraudulent patterns and label scarcity, this paper presents a novel model, Burstiness-aware Bipartite Graph Neural Networks (BurstBGN), which combats fraud by exploiting user-product bipartite graphs and timestamped rating activities. BurstBGN encapsulates two key ideas: the modeling of user-product interaction via historical rating data through an Edge-time GNN module and the exhaustive mapping of bursty fraudulent user activities. The performance of BurstBGN is demonstrated through rigorous benchmarking against established methods across three datasets. Our results show that BurstBGN consistently outperforms these methods under both transductive and inductive settings, confirming its effectiveness in detecting fraudulent users from limited annotated data, and thereby providing a safeguard for maintaining user trust in e-commerce platforms.

13 May 2024
Verified
Conference Paper
Open access
Cheng-Te LiCheng-Te Li
Cheng-Te Li
LKLun-Wei Ku
Lun-Wei Ku
YTYu-Che Tsai
Yu-Che Tsai
The field of SocialNLP stands at the confluence of two pivotal domains: natural language processing (NLP) and social computing, offering a multidisciplinary framework for exploring the intersections between these areas. This domain is characterized by its tri-directional focus. First, it endeavors to tackle prevalent challenges within the realm of social computing by leveraging advanced NLP methodologies. Second, it aims to address traditional NLP problems by harnessing the vast and dynamic data available from social networks and social media platforms. Third, SocialNLP is dedicated to identifying and solving emerging issues that lie at the nexus of social computing and NLP. The 12th iteration of the SocialNLP workshop, a hallmark event in this vibrant field, is scheduled to convene at TheWebConf (WWW) 2024. This edition has witnessed the acceptance of nine rigorously peer-reviewed papers, chosen from a competitive pool, reflecting an acceptance ratio of 53.8%. This achievement underscores the high caliber of research and the innovative approaches that the submitted works propose to advance the field of SocialNLP. The organizing committee extends its heartfelt gratitude to all contributing authors for their submissions, to the members of the program committee for their indispensable role in the rigorous review process, and to the workshop chairs for their visionary leadership and dedication. Their collective efforts have been instrumental in fostering an environment of academic excellence and collaboration that continues to drive forward the frontiers of knowledge in the interdisciplinary area of SocialNLP.