CIKM 2026

By Fabrice Popineau November 7, 2026

Two papers from the Lusis chair have been accepted at CIKM 2026, the 35th ACM International Conference on Information and Knowledge Management. The event will take place in Rome, Italy, on November 7–11, 2026. The main conference will run on November 9–11 at the Auditorium Parco della Musica, following tutorials and workshops on November 7–8.

• Mitigating Convergence Collapse in Fixed-Target Anomaly Detectors via Kernel-Anchored Locality Regularization, by José Lucas De Melo Costa, Fabrice Popineau, Arpad Rimmel and Bich-Liên Doan, is accepted for an oral presentation. The paper examines how prolonged training can erase the signal used by neural anomaly detectors. It introduces a kernel-based locality constraint to help preserve this signal.
• Knowledge-Informed Local Causal Discovery of Optimal Adjustment Sets, by Seong Woo Ahn, Alessandro Leite, José Lucas De Melo Costa, Fabrice Popineau, Bich-Liên Doan and Arpad Rimmel. The proposed b-LOAD algorithm incorporates background knowledge into local causal discovery to identify adjustment sets and improve causal effect estimation, particularly when data are limited.