
This paper introduces PUB, a novel multi-task optimization algorithm that addresses the "seesaw problem" in multi-task learning by balancing parameter updates instead of gradients, leading to improved performance and robustness across various applications, including recommendation systems and computer vision.


coremsg

parameter-update-balancing-pub-a-novel-multi-task-optimization-algorithm-for-recommendation-systems-and-beyond


Parameter Update Balancing (PUB): A Novel Multi-task Optimization Algorithm for Recommendation Systems and Beyond


title_rewrite


Aligning learning progress across tasks through adaptive group-based task interactions to improve overall multi-task performance while maintaining computational efficiency.


efficient-multi-task-optimization-through-adaptive-group-based-task-alignment


Efficient Multi-Task Optimization through Adaptive Group-Based Task Alignment



CoTBal algorithm enhances multi-task visual instruction tuning by balancing task contributions and difficulties.


cotbal-comprehensive-task-balancing-for-multi-task-visual-instruction-tuning


CoTBal: Comprehensive Task Balancing for Multi-Task Visual Instruction Tuning
