Our papers was accpeted at CIKM 2025

My PhD project paper, Toward Robust Machine Learning under Diverse Incomplete Data Mechanisms in Real-World Applications, was published in the Proceedings of the 34th ACM International Conference on Information and Knowledge Management (CIKM 2025).

The work outlines a systematic research agenda for developing machine learning methods that remain reliable under diverse incomplete data mechanisms, including MCAR, MAR, and MNAR, across multiple data modalities such as tabular, time-series, sensor, image, and textual data.

Our paper, MissDDIM: Deterministic and Efficient Conditional Diffusion for Tabular Data Imputation, was published in the Proceedings of the 34th ACM International Conference on Information and Knowledge Management (CIKM 2025).

The paper introduces a deterministic diffusion-based framework for tabular data imputation, addressing the high inference latency and output variability of existing stochastic diffusion models. MissDDIM adapts DDIM sampling to enable faster and more stable imputations while effectively leveraging incomplete inputs during training.