04:00
2026-06-17
arxiv.org
machine-learning
Pareto LoRA: Mitigating Modality Imbalance in Unified Multimodal Models via Pareto-Optimal Gradient Integration
Researchers from multiple institutions introduced Pareto LoRA, a method that mitigates modality imbalance in unified multimodal models by treating instruction tuning as a bi-objective optimization pro…