Hyunseung (Daniel) Hwang

aguno94 [at] gmail.com

I’m an incoming postdoctoral researcher at the Center for Responsible AI at New York University, working with Julia Stoyanovich. I recently completed my Ph.D. in Electrical Engineering at KAIST, advised by Steven Euijong Whang in the Data Intelligence Lab.

My research is on the interpretability and trustworthiness of machine learning. I stress-test the explanation methods that practitioners rely on to audit high-stakes models, and I show where they quietly break down. My recent work demonstrates that widely used tools such as SHAP are sensitive both to upstream data-engineering choices and to their own internal randomness — meaning the same model can yield materially different explanations. This has direct consequences for fairness, accountability, and whether an explanation can be trusted at all.

Earlier, I introduced XClusters, a framework that treats explainability as a first-class objective of clustering itself, rather than a step bolted on after the fact.

I’m broadly interested in responsible AI, model transparency, and the robustness of AI systems, and especially in turning these findings into something practitioners can act on — wherever models drive high-stakes decisions, from AI labs building trustworthy systems to financial services, where model risk, credit, and fair lending hinge on explanations that hold up. I’m open to both research and applied roles in industry where trustworthy AI is a priority.

selected publications

FAccT
SHAP-based Explanations are Sensitive to Feature Representation

Hyunseung Hwang, Andrew Bell, Joao Fonseca, and 3 more authors

In Proceedings of the 2025 ACM Conference on Fairness, Accountability, and Transparency (FAccT), 2025

Abs DOI arXiv Bib Code

Local feature-based explanations such as SHAP are central to validating, documenting, and contesting high-stakes models, including the credit and lending classifiers used throughout financial services. We show that routine data-engineering choices, such as how a continuous feature like age is bucketed or how race is encoded, can substantially shift SHAP feature-importance rankings for the same model, and that this can be exploited to make a discriminatory model appear neutral. The findings bear directly on model risk management, fair-lending and adverse-action compliance, and whether a feature-importance explanation can be trusted at all.
@inproceedings{hwang2025shap, title = {SHAP-based Explanations are Sensitive to Feature Representation}, author = {Hwang, Hyunseung and Bell, Andrew and Fonseca, Joao and Pliatsika, Venetia and Stoyanovich, Julia and Whang, Steven Euijong}, booktitle = {Proceedings of the 2025 ACM Conference on Fairness, Accountability, and Transparency (FAccT)}, pages = {1588--1601}, year = {2025}, doi = {10.1145/3715275.3732105}, }
AAAI
XClusters: Explainability-first Clustering

Hyunseung Hwang and Steven Euijong Whang

In Proceedings of the AAAI Conference on Artificial Intelligence (AAAI), 2023

Abs arXiv Bib

Most explainable-clustering methods explain clusters only after the fact. XClusters instead makes explainability a first-class objective of the clustering itself, jointly minimizing cluster distortion and the size of the decision tree that explains the clusters, which matters wherever segments must be both accurate and defensible, as in customer and credit-risk segmentation (we evaluate on a real credit-card dataset). We show the objective is a difference of monotonic functions and give an efficient branch-and-bound algorithm that trades off accuracy against explanation simplicity.
@inproceedings{hwang2023xclusters, title = {XClusters: Explainability-first Clustering}, author = {Hwang, Hyunseung and Whang, Steven Euijong}, booktitle = {Proceedings of the AAAI Conference on Artificial Intelligence (AAAI)}, year = {2023}, }