1 paper · 1 filter
Mingfei Han, Liang Ma, Kamila Zhumakhanova +5
Vision-and-Language Navigation (VLN) suffers from the limited diversity and scale of training data, primarily constrained by the manual curation of existing simulators. To address…