不重训练，用动态记忆衰减和防循环惩罚提升视觉导航效率
DART-VLN: Test-Time Memory Decay and Anti-Loop Regularization for Discrete Vision-Language Navigation
- 测试时引入记忆衰减，自动降低过时记忆权重
- 加入反循环惩罚，减少原地折返，路径更短
- 无需新增参数，兼容现有模型，适合部署优化
基于记忆的离散视觉语言导航（VLN）代理在部分可观测环境下运行，即使使用强大预训练骨干网络，仍可能产生系统性推理错误。本文聚焦两个常见问题：记忆读取时的历史证据过时、动作选择时局部回溯效率低。提出 DART-VLN，一种无需训练的推理时框架，结合测试时记忆衰减（重加权过时冗余记忆槽，不修改存储内容）与反循环正则化（轻量级下一步惩罚，抑制立即反向动作）。该方法不引入可学习参数，保持导航骨干不变。在 R2R 与 REVERIE 数据集上的实验表明，记忆衰减能持续维持或提升任务性能的同时减少运行时间；加入反循环正则化后进一步缩短路径，减少局部回溯，在对比的 GridMM 变体中实现导航质量与效率的最佳平衡。结果表明，轻量级推理时控制可提升记忆型离散 VLN 的可靠性和效率，无需重新训练。
原文摘要 · Abstract (English)
Memory-based agents for discrete vision-language navigation (VLN) operate under partial observability and can exhibit systematic inference-time failures even with strong pretrained backbones. We focus on two recurring problems: stale historical evidence during memory readout and inefficient local backtracking during action selection. We present DART-VLN, a training-free inference-time framework that combines Test-Time Memory Decay, which reweights stale and redundant memory slots without modifying their stored content, with Anti-Loop Regularization, a lightweight next-hop penalty that discourages immediate reversals. DART-VLN introduces no learnable parameters and leaves the navigation backbone unchanged. Experiments on R2R and REVERIE show that memory decay consistently preserves or improves task performance while reducing runtime. Adding anti-loop regularization further shortens trajectories, reduces local backtracking, and achieves the best overall balance between navigation quality and efficiency among the evaluated GridMM variants. These results indicate that lightweight inference-time control can improve the reliability and efficiency of memory-based discrete VLN without retraining.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。