Edge-aware decoder exposes transition-level quantities at decision time, cutting the ATSP-1000 gap from 4.13% to 2.73% over RADAR on SVD/Sinkhorn backbone.
ASP: Learn a universal neural solver!IEEE Transactions on Pattern Analysis and Machine Intelligence, 46(6):4102– 4114, 2024a
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.LG 2years
2026 2verdicts
UNVERDICTED 2representative citing papers
DREE enables neural VRP solvers to learn new drifting tasks efficiently while preserving prior knowledge and improving generalization under limited per-task training resources.
citing papers explorer
-
Edge-aware Decoding for Neural Asymmetric Routing
Edge-aware decoder exposes transition-level quantities at decision time, cutting the ATSP-1000 gap from 4.13% to 2.73% over RADAR on SVD/Sinkhorn backbone.
-
Keep Rehearsing and Refining: Lifelong Learning Vehicle Routing under Continually Drifting Tasks
DREE enables neural VRP solvers to learn new drifting tasks efficiently while preserving prior knowledge and improving generalization under limited per-task training resources.