Research map / Vision-language and multimodal / Robot action and navigation
HUGE-Bench: A Benchmark for High-Level UAV Vision-Language-Action Tasks
Jingyu Guo, Ziye Chen, Ziwen Li, Zhengqing Gao, Jiaxin Huang, Hanlue Zhang, Fengming Huang, Yu Yao, Tongliang Liu, Mingming Gong
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add implicit memory to OpenVLA (MemoryVLA) | OpenVLA · HUGE-Bench test (high-level instructions) | Avg. TCR: 0.112 -> 0.234 | Improved |
| add implicit memory to OpenVLA (MemoryVLA) | OpenVLA · HUGE-Bench test (high-level instructions) | NTP: 0.122 -> 0.549 | Improved |
| add depth maps as extra visual input to pi_0.5 | pi_0.5 · HUGE-Bench test (high-level instructions) | CR: 0.018 -> 0.013 | Improved |
| add depth maps as extra visual input to pi_0.5 | pi_0.5 · HUGE-Bench test (high-level instructions) | nDTW: 0.467 -> 0.459 | Hurt |
| add depth maps as extra visual input to pi_0.5 | pi_0.5 · HUGE-Bench test (high-level instructions) | NTP: 0.618 -> 0.617 | No clear effect |
| use high-level instead of low-level instructions | pi_0.5 · HUGE-Bench test (high-level instructions) | Avg. TCR: 0.698 -> 0.581 | Hurt |
| use high-level instead of low-level instructions | FastVLM · HUGE-Bench test (high-level instructions) | Avg. TCR: 0.476 -> 0.285 | Hurt |
| use high-level instead of low-level instructions | pi_0 · HUGE-Bench test (high-level instructions) | nDTW: 0.561 -> 0.443 | Hurt |
| use high-level instead of low-level instructions | OpenVLA · HUGE-Bench test (high-level instructions) | CR: 0.008 -> 0.003 | Mixed |
| use pi_0.5 instead of pi_0 | pi_0 policy · HUGE-Bench test (high-level instructions) | Avg. TCR: 0.558 -> 0.581 | Improved |
| use pi_0.5 instead of pi_0 | pi_0 policy · HUGE-Bench test (high-level instructions) | CR: 0.011 -> 0.018 | Hurt |
| add explicit frame-history grids from past 20 or 50 meters to OpenVLA-style baseline (ablation) | OpenVLA/MemoryVLA · HUGE-Bench | task performance | Hurt |
Closest papers by meaning
- Towards Realistic UAV Vision-Language Navigation: Platform, Benchmark, and Methodology (ICLR 2025 · similarity 0.87)
- InfiniBench: Infinite Benchmarking for Visual Spatial Reasoning with Customizable Scene Complexity (CVPR 2026 · similarity 0.86)
- QUANTIPHY: A Quantitative Benchmark Evaluating Physical Reasoning Abilities of Vision-Language Models (CVPR 2026 · similarity 0.86)
- AirCopBench: A Benchmark for Multi-drone Collaborative Embodied Perception and Reasoning (AAAI 2026 · similarity 0.86)
- Is your VLM Sky-Ready? A Comprehensive Spatial Intelligence Benchmark for UAV Navigation (CVPR 2026 · similarity 0.85)
- SatNav: A Scalable Benchmark for Long-Horizon UAV Vision-Language Navigation from Satellite Imagery (NeurIPS 2026 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
