DivMerge: A divergence-based model merging method for multi-tasking
This addresses the problem of scaling model merging for multi-tasking in machine learning, though it appears incremental as it builds on existing model merging approaches.
The paper tackles task interference in multi-task learning by merging models trained on different tasks into a single model, achieving strong performance across all tasks without needing additional labeled data.
Multi-task learning (MTL) is often achieved by merging datasets before fine-tuning, but the growing availability of fine-tuned models has led to new approaches such as model merging via task arithmetic. A major challenge in this setting is task interference, which worsens as the number of tasks increases. We propose a method that merges models trained on different tasks into a single model, maintaining strong performance across all tasks. Our approach leverages Jensen-Shannon divergence to guide the merging process without requiring additional labelled data, and automatically balances task importance. Unlike existing methods, our approach remains robust as the number of tasks grows and consistently outperforms prior work.