Entrar
insight - Multi-Task Optimization
Parameter Update Balancing (PUB): A Novel Multi-task Optimization Algorithm for Recommendation Systems and Beyond
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.
Efficient Multi-Task Optimization through Adaptive Group-Based Task Alignment
Aligning learning progress across tasks through adaptive group-based task interactions to improve overall multi-task performance while maintaining computational efficiency.
CoTBal: Comprehensive Task Balancing for Multi-Task Visual Instruction Tuning
CoTBal algorithm enhances multi-task visual instruction tuning by balancing task contributions and difficulties.
Produtos
© 2024 by Linnk AI