---
title: "Georgia Tech's FiLM-Nav Teaches Robots to Navigate Any Room 20% More Efficiently"
subtitle: "Georgia Tech researchers fine-tune a Mamba-based vision-language model as a frontier picker, topping HM3D ObjectNav and HM3D-OVON on path efficiency."
url: "https://alphasignal.ai/news/georgia-tech-s-film-nav-teaches-robots-to-navigate-any-room-20-more-efficiently"
author: "AlphaSignal Newsroom"
author_url: "https://alphasignal.ai/editorial-team"
company: "AlphaSignal"
company_url: "https://alphasignal.ai/company/alphasignal"
published: "2026-04-15T00:00:00.000Z"
updated: "2026-08-23T21:37:12.454Z"
type: "Paper"
topics:
  - "Robotics"
  - "Post Training"
subtopics:
  - "Vla Models"
  - "Fine Tuning"
upvotes: 2702
source_url: "https://arxiv.org/abs/2509.16445"
comments: 0
truncated: true
---

# Georgia Tech's FiLM-Nav Teaches Robots to Navigate Any Room 20% More Efficiently

> Georgia Tech researchers fine-tune a Mamba-based vision-language model as a frontier picker, topping HM3D ObjectNav and HM3D-OVON on path efficiency.

Getting a robot to find the "gravity chair" in a house it has never seen sounds simple until you try it. The robot needs to understand what a gravity chair is, remember which rooms it has already walked through, and decide which unexplored doorway to try next. A new paper called [FiLM-Nav](https://arxiv.org/abs/2509.16445) from Georgia Tech tackles this by taking a pre-trained vision-language model and fine-tuning it directly as the navigation policy, rather than using it as a bolt-on reasoning module.

[FiLM-Nav](https://filmnav.netlify.app) sets a new state of the art on both the HM3D ObjectNav and the harder HM3D-OVON open-vocabulary benchmark, especially on path efficiency, the metric that matters when a robot has finite battery.

## Why plugging a VLM into a robot is harder than it looks

Object Goal Navigation (ObjectNav) drops an agent in an unseen 3D environment and tells it, in language, to go find something. The open-vocabulary version, [OVON](https://arxiv.org/abs/2409.14296), ups the difficulty by allowing any object description in free-form language, rather than a fixed list of six categories.

Existing approaches often use foundation models for high-level planning, instruction interpretation, reasoning about object locations from semantic cues, or enriching environmental representations. Because these models are typically trained on disembodied web data, they exhibit weaknesses in spatial-temporal reasoning and physical grounding, both critical for efficient robotic task execution. Common pipelines pass a text summary of what the camera saw into an LLM, ask for a plan, and lose most of the visual detail in the translation.

_Preview only. Full text requires an AlphaSignal Pro account._

## Sources

- [FiLM-Nav: Efficient and Generalizable Navigation via VLM Fine-tuning](https://arxiv.org/abs/2509.16445)
- [FiLM-Nav Project Page](https://filmnav.netlify.app)
- [Cobra: Extending Mamba to Multi-Modal Large Language Model](https://arxiv.org/abs/2403.14520)
- [HM3D-OVON: A Dataset and Benchmark for Open-Vocabulary Object Goal Navigation](https://arxiv.org/abs/2409.14296)