Xiaoyang Wang

Mr. Xiaoyang Wang

orcid

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

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

Representative Publications

View All Publications
13 August 2025 in IEEE Transactions on Information Forensics and Security
Verified
Journal Article
Closed access
YLYixuan Li
Yixuan Li
XLXuelin Liu
Xuelin Liu
XWXiaoyang Wang
Xiaoyang Wang
BLBu Sung Lee
Bu Sung Lee
SWShiqi Wang
Shiqi Wang
WLWeisi Lin
Weisi Lin
The ability to distinguish whether an image is generated by artificial intelligence (AI) is a crucial ingredient in human intelligence, usually accompanied by a complex and dialectical forensic and reasoning process. However, current fake image detection models and databases focus on binary classification without understandable explanations for the general populace. This weakens the credibility of authenticity judgment and may conceal potential model biases. Meanwhile, large multimodal models (LMMs) have exhibited immense vision-language capabilities on various tasks, bringing the potential for explainable fake image detection. Therefore, we pioneer the probe of LMMs for explainable fake image detection by presenting a multimodal database encompassing descriptions of textual authenticity, the FakeBench. For construction, we first introduce a fine-grained taxonomy of generative visual forgery concerning human perception, based on which we collect forgery descriptions in human natural language with a human-in-the-loop strategy. FakeBench examines LMMs with four evaluation criteria: detection, reasoning, explanation and fine-grained forgery analysis, to obtain deeper insights into image authenticity-relevant capabilities. Experiments on various LMMs confirm their merits and demerits in different aspects of fake image detection tasks. This research presents a paradigm shift towards transparency for the fake image detection area and reveals the need for greater emphasis on forensic elements in visual-language research and AI risk control. FakeBench will be available at https://github.com/Yixuan423/FakeBench


25 July 2024
Verified
Book Chapter
Closed access
PDF file
CCChia-Hsuan Chang
Chia-Hsuan Chang
XWXiaoyang Wang
Xiaoyang Wang
CYChristopher C Yang
Christopher C Yang


29 August 2025 in arXiv
Verified
Preprint
Open access
PDF file
YLYeawon Lee
Yeawon Lee
XWXiaoyang Wang
Xiaoyang Wang
CYChristopher C Yang
Christopher C Yang
Accurate interpretation of clinical narratives is critical for patient care, but the complexity of these notes makes automation challenging. While Large Language Models (LLMs) show promise, single-model approaches can lack the robustness required for high-stakes clinical tasks. We introduce a collaborative multi-agent system (MAS) that models a clinical consultation team to address this gap. The system is tasked with identifying clinical problems by analyzing only the Subjective (S) and Objective (O) sections of SOAP notes, simulating the diagnostic reasoning process of synthesizing raw data into an assessment. A Manager agent orchestrates a dynamically assigned team of specialist agents who engage in a hierarchical, iterative debate to reach a consensus. We evaluated our MAS against a single-agent baseline on a curated dataset of 420 MIMIC-III notes. The dynamic multi-agent configuration demonstrated consistently improved performance in identifying congestive heart failure, acute kidney injury, and sepsis. Qualitative analysis of the agent debates reveals that this structure effectively surfaces and weighs conflicting evidence, though it can occasionally be susceptible to groupthink. By modeling a clinical team's reasoning process, our system offers a promising path toward more accurate, robust, and interpretable clinical decision support tools.

17 December 2024 in IEEE Access
Verified
Journal Article
Open access
XWXiaoyang Wang
Xiaoyang Wang
CCChia-Hsuan Chang
Chia-Hsuan Chang
CYChristopher C Yang
Christopher C Yang
Machine learning algorithms are increasingly used in real-world decision-making systems, raising concerns about potential biases and unfairness. Existing in-processing bias mitigation approaches often focus on achieving numerical parity across demographic groups while neglecting the performance impact on individual groups and the overall performance-fairness trade-off. This can lead to increased discriminatory outcomes and hinder efforts to develop truly fair AI systems. This paper proposes Transfer Learning with Fairness Optimization (TLFO), a novel framework that serially optimizes predictive performance and fairness in machine learning models. TLFO leverages transfer learning by dividing the training process into two distinct phases: (1) initial learning for performance optimization and (2) subsequent finetuning for fairness enhancement. This sequential approach enables fine-grained control over fairness constraints, minimizing the performance-fairness trade-off. Extensive experiments on two real-world datasets demonstrate TLFO’s effectiveness. TLFO consistently achieves superior fairness with minimal performance degradation compared to state-of-the-art in-processing bias mitigation approaches, highlighting its potential for generating fair and accurate classifiers with versatile applications.