MS student, Southern University of Science and Technology
4 papers at NeurIPS 2025
This paper proposes the Multi-Task Learning with Knowledge Distillation (MTL-KD) to train a single heavy decoder model without labeled data to solve various VRP variants.