FreqCache: Accelerating Embodied VLN Models with Adaptive Frequency-Guided Token Caching

2026-04-27

AI Summary

This work addresses the limitations of existing token caching methods in vision-and-language navigation (VLN), which struggle to balance efficiency and performance due to viewpoint variations, loss of critical edge information, and rigid cache budgets. To overcome these challenges, we introduce frequency-domain analysis into VLN token caching for the first time, leveraging its viewpoint invariance and structural interpretability. We propose a training-free, adaptive caching framework that dynamically optimizes cache construction, refreshment, and budget allocation. This approach effectively preserves essential visual edge features and mitigates viewpoint shift effects, achieving a 1.59× inference speedup with negligible computational overhead, thereby significantly enhancing the inference efficiency of VLN models.

Technology Category

• Computer Vision: Language and Vision
• Planning, Routing, and Scheduling: Planning with Language Models
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines

Abstract

Vision-Language-Navigation (VLN) models exhibit excellent navigation accuracy but incur high computational overhead. Token caching has emerged as a promising training-free strategy to reduce this cost by reusing token computation results; however, existing token caching approaches rely on visual domain methods for cacheable token selection, leading to challenges when adapted to VLN models. 1) Visual domain methods become invalid when there is viewpoint migration. 2) Visual domain methods neglect critical edge information without the aid of additional algorithms. 3) Visual domain methods overlook the temporal variation of scenarios and lack adjustability in cache budgets. In this paper, we develop detailed analyses and find that the impacts of these challenges exhibit invariance and analyzability in the frequency domain. Based on these, we propose a frequency-guided token caching framework, called FreqCache. Utilizing the inherent properties of the frequency domain, FreqCache achieves optimal token cache establishment, refreshment, and adaptive adjustment. Experiments show that FreqCache achieves 1.59x speedup with ignorable overhead, showing the effect of integrating frequency domain methods in VLN token caching.

Problem

• token caching
• Vision-Language Navigation
• frequency domain
• computational overhead
• viewpoint migration

Innovation

• frequency domain
• token caching
• embodied navigation
• adaptive caching
• Vision-Language Navigation