Task Affinity

Task affinity in multi-task learning (MTL) focuses on identifying relationships between different tasks to improve the efficiency and effectiveness of training models that perform multiple tasks simultaneously. Current research emphasizes developing efficient algorithms to estimate task affinities, including those capturing higher-order relationships beyond pairwise comparisons, and using these affinities to optimize task grouping and model architectures, such as dynamically adjusting network structures based on task importance and resource constraints. This research is significant because it addresses the challenges of negative transfer in MTL—where some tasks hinder others—leading to improved performance and reduced computational costs across diverse applications like graph neural networks and image processing.