我所汪群博老师参与完成的论文《FlexVLN: Flexible Adaptation for Diverse Vision-and-Language Navigation Tasks》于2025-09在《IEEE Transactions on Multimedia》正式发表。论文摘要如下：
Abstract: The aspiration of the Vision-and-Language Navigation (VLN) task has long been to develop an embodied agent with robust adaptability, capable of seamlessly transferring its navigation capabilities across various tasks. Despite remarkable advancements in recent years, most methods necessitate dataset-specific training, thereby lacking the capability to generalize across diverse datasets encompassing distinct types of instructions. Large language models (LLMs) have demonstrated exceptional reasoning and generalization abilities, exhibiting immense potential in robot action planning. In this paper, we propose FlexVLN, an innovative hierarchical approach to VLN that integrates the fundamental navigation ability of a supervised-learning-based Instruction Follower with the robust generalization ability of the LLM Planner, enabling effective generalization across diverse VLN datasets. Moreover, a verification mechanism and a multi-model integration mechanism are proposed to mitigate potential hallucinations by the LLM Planner and enhance execution accuracy of the Instruction Follower. We take REVERIE, SOON, and CVDN-target as out-of-domain datasets for assessing generalization ability. The generalization performance of FlexVLN surpasses that of all the previous methods to a large extent.
引用信息：Siqi Zhang, Yanyuan Qiao, Qunbo Wang, Longteng Guo, Zhihua Wei, Jing Liu. FlexVLN: Flexible Adaptation for Diverse Vision-and-Language Navigation Tasks. IEEE Transactions on Multimedia, 2025, 27: 6307-6318. DOI: 10.1109/TMM.2025.3581809