TRAVEL: Training-Free Retrieval and Alignment for Vision-and-Language Navigation

TL;DR

This paper introduces a modular, training-free approach for vision-and-language navigation that leverages large language and vision-language models in a zero-shot setting to improve path prediction and alignment accuracy.

Contribution

It proposes a novel modular framework that decomposes VLN into sub-tasks using LLMs and VLMs without training, enhancing navigation performance on complex datasets.

Findings

Outperforms joint semantic map-based methods on R2R-Habitat dataset

Effectively extracts landmarks and generates path hypotheses

Achieves superior path alignment and navigation accuracy

Abstract

In this work, we propose a modular approach for the Vision-Language Navigation (VLN) task by decomposing the problem into four sub-modules that use state-of-the-art Large Language Models (LLMs) and Vision-Language Models (VLMs) in a zero-shot setting. Given navigation instruction in natural language, we first prompt LLM to extract the landmarks and the order in which they are visited. Assuming the known model of the environment, we retrieve the top-k locations of the last landmark and generate $k$ path hypotheses from the starting location to the last landmark using the shortest path algorithm on the topological map of the environment. Each path hypothesis is represented by a sequence of panoramas. We then use dynamic programming to compute the alignment score between the sequence of panoramas and the sequence of landmark names, which match scores obtained from VLM. Finally, we compute the nDTW metric between the hypothesis that yields the highest alignment score to evaluate the path fidelity. We demonstrate superior performance compared to other approaches that use joint semantic maps like VLMaps on the complex R2R-Habitat instruction dataset and quantify in detail the effect of visual grounding on navigation performance.