We are excited to share that the 43rd International Conference on Machine Learning (ICML 2026) in Seoul will feature four papers from the AImageLab team. Congratulations to all the authors!
– “Distilling Linearized Behavior into Non-linear Fine-tuning for Effective Task Arithmetic”, by Thomas Sommariva, Francesca Morandi, Simone Calderara, and Angelo Porrello. (Paper)
– “Restoring Exploration after Post-Training: Latent Exploration Decoding for Large Reasoning Models” by Wenhui Tan, Fiorenzo Parascandolo, Enver Sangineto, Jianzhong Ju, Zhenbo Luo, Qian Cao, Rita Cucchiara, Ruihua Song, Jian Luan. (Paper)
– “Shifting the Breaking Point of Flow Matching for Multi-Instance Editing” by Carmine Zaccagnino, Fabio Quattrini, Enis Simsar, Marta Tintoré Gazulla, Rita Cucchiara, Alessio Tonioni, and Silvia Cascianelli. (Paper)
– “Transporting Task Vectors Across Different Architectures Without Training” by Filippo Rinaldi, Aniello Panariello, Giacomo Salici, Angelo Porrello, and Simone Calderara (Paper)
Big shoutout to Filippo, Silvia, and Fiorenzo, who were recognized as Gold, Gold, and Silver Reviewers, respectively!
We look forward to presenting our work and meeting the research community in Seoul. See you at the COEX Convention Center from July 6 to 11!