Research map / Image generation / Fast diffusion sampling
GeniNav: Generative Model Driven Image-Goal Navigation via Imagination-Guided Consistency Flow Matching
Yuqi Chen, Junjie Gao, Yongzhou Pan, Siyuan Song, Zixuan Zhang, Jiaping Xiao, Mir Feroskhan
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 latent guidance module (LGM) for subgoal imagination (ablation) | GeniNav · GeniBench Gibson | SR (%): 58.2 -> 68.7 | Improved |
| add auxiliary semantic, geometric and contrastive losses to LGM (ablation) | GeniNav · GeniBench Gibson | SR (%): 61.8 -> 68.7 | Improved |
| use latent subgoal instead of explicit diffusion image subgoal (ablation) | GeniNav · GeniBench Gibson | SR (%): 62.5 -> 68.7 | Improved |
| use multi-segment consistency flow matching instead of single-field CFM (ablation) | GeniPolicy · GeniBench Gibson | SR (%): 63.7 -> 68.7 | Improved |
| use MS-CFM instead of diffusion policy under same inference budget (ablation) | GeniPolicy · GeniBench Gibson | SR (%): 39.0 -> 68.7 | Improved |
| use hybrid ranking module instead of random trajectory selection (ablation) | GeniNav · GeniBench Gibson | SR (%): 53.2 -> 68.7 | Improved |
| use HRM instead of critic-based evaluator (ablation) | GeniNav · GeniBench Gibson | SR (%): 63.8 -> 68.7 | Improved |
| use HRM (semantic plus geometric) instead of VLM-only evaluator (ablation) | GeniNav · GeniBench Gibson | SR (%): 60.3 -> 68.7 | Improved |
| add latent guidance module (ablation) | GeniNav · GeniBench MP3D (cross-domain) | SR (%): 43.3 -> 55.2 | Improved |
| use GeniNav instead of strongest baseline MetricNet | image-goal navigation · GeniBench Gibson | SR / SPL (%): SR +14.2, SPL +16.1 | Improved |
Closest papers by meaning
- ImagineNav: Prompting Vision-Language Models as Embodied Navigator through Scene Imagination (ICLR 2025 · similarity 0.88)
- GoalFlow: Goal-Driven Flow Matching for Multimodal Trajectories Generation in End-to-End Autonomous Driving (CVPR 2025 · similarity 0.87)
- Distilling LLM Prior to Flow Model for Generalizable Agent’s Imagination in Object Goal Navigation (NeurIPS 2025 · similarity 0.87)
- LatentPilot: Scene-Aware Vision-and-Language Navigation by Dreaming Ahead with Latent Visual Reasoning (ECCV 2026 · similarity 0.87)
- Latent Barrier Steering: Hierarchical Safety for Generative Planning (NeurIPS 2026 · similarity 0.87)
- Generative Language-Grounded Policy in Vision-and-Language Navigation with Bayes' Rule (ICLR 2021 · similarity 0.87)
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.
