Publications

You can also find my articles on my Google Scholar profile.

Locality-aware Diffusion Language Modeling

Published in arXiv preprint, 2026, 2026

Studies the trainability of Masked Diffusion Language Models against autoregressive LLMs across structured generation tasks. Proposes two locality-aware blockwise diffusion architectures (Scatter and Jigsaw) bridging AR-style local ordering and Diffusion-style global iterative refinement.

Recommended citation: Yuxiang Wang, et al. "Locality-aware Diffusion Language Modeling." arXiv:2604.24832, 2026.
Download Paper

Semantic Diffusion Language Modeling

Published in Preprint, 2026 (Second author), 2026

Investigates how the noising-kernel design in discrete diffusion language models affects training stability and generation quality. Proposes SemDLM with semantic-neighborhood diffusion, a shared refresh branch, and token-frequency prior correction. Achieves 27.19 Test PPL on LM1B.

STM: Spatio-Temporal Distance and Frame-Based Dynamic Graph Fraud Detection

Published in CCF-B venue (Springer), 2024

Proposes STM, a dynamic-graph fraud-detection model that combines graph condensation, distance-aware intra-frame aggregation, and difference-aware inter-frame aggregation to fully exploit short- and long-range spatio-temporal information. Achieves SOTA on multiple dynamic-graph benchmarks. Also resulted in an invention patent.

Recommended citation: Yuxiang Wang, et al. "Spatio-Temporal Distance and Frame-Based Dynamic Graph Fraud Detection." Springer, 2024. https://doi.org/10.1007/978-981-95-3906-2_3
Download Paper