Pith. sign in

REVIEW 4 major objections 4 minor 7 cited by

DreamNav claims that egocentric-only, trajectory-planning zero-shot navigation with imagined future narration outperforms all panoramic and odometry-aided baselines on R2R-CE Val-Unseen.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · deepseek-v4-flash

2026-08-04 16:55 UTC pith:3T7KD7VT

load-bearing objection The strongest zero-shot egocentric VLN claim needs a cleaner evaluation protocol before it's convincing; the system design is plausible and worth a referee. the 4 major comments →

arxiv 2509.11197 v1 pith:3T7KD7VT submitted 2025-09-14 cs.RO cs.AIcs.CLcs.CV

DreamNav: A Trajectory-Based Imaginative Framework for Zero-Shot Vision-and-Language Navigation

Yunheng Wang , Yuetong Fang , Taowen Wang , Yixiao Feng , Yawen Tan , Shuning Zhang , Peiran Liu , Yiding Ji
show 1 more author
Renjing Xu
This is my paper
classification cs.RO cs.AIcs.CLcs.CV
keywords vision-and-language navigationzero-shot navigationegocentric perceptiontrajectory planningimaginationdiffusion policyworld modelVLN-CE
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper sets out to show that a zero-shot vision-and-language navigation agent can operate from a single egocentric RGB-D camera, without panoramic sensors or odometry, by planning whole trajectories rather than point-to-point hops and by 'imagining' what lies along each candidate route. It introduces DreamNav, whose EgoView Corrector stabilizes the egocentric view, whose Trajectory Predictor generates diverse navigable paths, and whose Imagination Predictor produces narrated future scenarios that a Navigator ranks against the instruction. On the R2R-CE Val-Unseen benchmark, the authors report SR 32.79% and SPL 28.95%, beating the strongest panoramic competitor and an egocentric baseline that also uses odometry by 7.49% SR and 18.15% SPL. The paper further shows the system transferring to a real robot, completing 12 of 20 indoor tasks where an open-source zero-shot baseline completes 6. The broader claim is that human-like thinking ahead plus trajectory-level semantics can substitute for expensive global perception in zero-shot navigation.

Core claim

The central discovery claimed is that, within the zero-shot VLN-CE setting, a purely egocentric agent that plans over whole trajectories and actively imagines their futures can outperform both panoramic-sensing methods and egocentric methods that use odometry. The reported numbers on R2R-CE Val-Unseen are success rate 32.79% and SPL 28.95%, compared with 31.00% SR and 24.00% SPL for the strongest panoramic method, and 25.30% SR and 10.80% SPL for the odometry-aided egocentric baseline. In real-world trials, DreamNav succeeds in 12 of 20 tasks versus 6 of 20 for an open-source zero-shot baseline and 3 of 20 for a supervised egocentric baseline. The authors attribute the gains to three compone

What carries the argument

The load-bearing mechanism is a closed loop of generation, imagination, and selection. A two-stage EgoView Corrector (a macro rotation expert plus a micro walkability-gated controller) keeps the egocentric camera aligned with the instruction. A diffusion-policy Trajectory Predictor samples 24-waypoint candidate paths and picks a diverse subset by farthest-first selection. The Imagination Predictor then feeds each candidate trajectory into a controllable world model that synthesizes egocentric future frames, and a narration model converts those frames into task-focused textual descriptions. Finally, the Navigator ranks the descriptions against the current subtask, while an Execution Expert tr

Load-bearing premise

The load-bearing premise is that the controllable world model generates coherent, geometrically faithful egocentric rollouts in unseen environments, and that the narration of those rollouts conveys enough task-relevant semantics to rank candidate trajectories correctly; if the imagined frames are mostly hallucination, the Navigator is effectively choosing on noise.

What would settle it

On R2R-CE Val-Unseen, replace the Imagination Predictor's output with (a) paraphrased random text and (b) ground-truth frames rendered from the simulator along each candidate trajectory, keeping everything else fixed. If random-text ranking matches DreamNav's SR, the gains are not from imagination; if ground-truth-frame ranking substantially exceeds it, the world model's fidelity is the limiting factor.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • Egocentric-only sensing is sufficient for competitive zero-shot VLN-CE, removing the need for panoramic capture or multi-sensor rigs at deployment.
  • Trajectory-level action selection aligns better with instruction semantics than point-level waypoint prediction, yielding higher SR and SPL.
  • Imagination helps: increasing imagined rollout length improves decisions up to a point (IRL=18), after which accumulated noise degrades performance.
  • The system transfers to a real robot, with 12/20 task successes versus 6/20 for a leading open-source zero-shot baseline.
  • Navigation error remains higher than panoramic methods, an acknowledged cost of losing global context.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The authors do not isolate how much of the gain comes from trajectory-level planning versus imagination; a clean test would keep the Navigator but rank trajectories using only path geometry plus the current view, without narrated rollouts.
  • If synthesized rollouts are partly hallucinated yet still help, the implication is that semantic plausibility matters more than geometric truth for instruction ranking; that could be tested by swapping the world model for a text-only 'imagine the scene' prompt.
  • The same trajectory-to-text imagination loop could extend to other long-horizon embodied tasks, such as object search or mobile manipulation, where a language model ranks candidate motion plans by narrated futures.
  • The IRL=18 optimum hints that uncertainty accumulates with rollout length; an adaptive horizon that truncates imagination when predicted confidence drops might improve on the fixed length.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

4 major / 4 minor

Summary. DreamNav proposes a zero-shot Vision-and-Language Navigation in Continuous Environments (VLN-CE) system that uses only monocular egocentric RGB-D input. The framework combines four modules: an EgoView Corrector that reduces orientation errors via a pretrained expert and a FastSAM-based micro-adjust controller; a Trajectory Predictor that generates candidate trajectories with a diffusion policy and prunes them via farthest-first selection; an Imagination Predictor that uses a controllable world model (Stable Virtual Camera) and Qwen-VL to produce textual descriptions of imagined rollouts; and a Navigation Manager implemented with GPT-4o that selects among trajectories and monitors subtask progress. The authors report state-of-the-art zero-shot results on R2R-CE Val-Unseen (Table I) and on a 20-trial real-world benchmark (Table II), claiming that trajectory-level planning and active imagination enable egocentric-only navigation to surpass panoramic and odometry-assisted baselines.

Significance. If the results hold, the paper makes a useful contribution: it shows that egocentric-only zero-shot VLN-CE is feasible and can outperform methods that rely on panoramic sensing or odometry, which would lower sensor cost and broaden deployability. The idea of converting imagined visual rollouts into textual narratives for an LLM-based navigator is also practically interesting and avoids the cost of passing raw generated frames to a vision-language model. The modular architecture is clearly described, and the component ablations (Table III, Figure 5) show large effects for the EgoView Corrector and for the choice of imagination rollout length. However, the empirical evidence is currently undermined by the evaluation protocol: hyperparameters are selected on a subset of the same Val-Unseen episodes used for the headline numbers, and no variance estimates are provided. The central SOTA claim therefore needs a stricter, held-out evaluation before it can be accepted.

major comments (4)
  1. [Sec. IV-A, Sec. IV-D, Fig. 5, Table I] The hyperparameters CTN=4 and IRL=18 are selected using ablations on 100 randomly chosen episodes from the Val-Unseen split, and the same split is then used for the final comparisons in Table I. This leaks evaluation information into model selection and can inflate the reported gains. The authors should either tune on a separate split (e.g., Val-Seen or a held-out portion of Val-Unseen) and report results only on the untouched episodes, or clearly show that the selected values are insensitive to the tuning split. As written, the headline 'surpasses all panoramic-based methods' claim is not supported by a clean evaluation protocol.
  2. [Table I and Sec. IV-B] No variance or confidence intervals are reported anywhere, despite the pipeline containing stochastic diffusion rollouts, GPT-4o and Qwen-VL API calls, and randomly selected evaluation episodes. The reported improvements over InstructNav are small (1.79% SR, 4.95% SPL); without error bars or multiple-seed results, these gaps may be within run-to-run variability. The authors should report at least bootstrap confidence intervals over episodes or multiple independent runs, especially for the CA-Nav comparison where the claimed gain is 7.49% SR and 18.15% SPL.
  3. [Sec. IV-C, Table II] The real-world evaluation consists of only 20 trials total. The claim that DreamNav 'surpasses Open-Nav by 30%' is based on 12/20 versus 6/20 successes; with this sample size, a few episode outcomes would change the conclusion. The paper should either provide more trials, report per-scene variability, or use a statistical test. As presented, the real-world superiority claim is not robustly supported.
  4. [Sec. III-D, Eq. (8), Fig. 5(b)] The Imagination Predictor is a core contribution, but its validity rests on the assumption that Stable Virtual Camera produces geometrically faithful visual rollouts and that Qwen-VL's narrated descriptions are semantically informative for trajectory ranking. The only direct evidence is the IRL=0 no-imagination ablation in Fig. 5(b), which is limited to 100 episodes and is part of the test-set tuning issue raised above. The paper would benefit from a sanity check of the imagined frames against ground-truth observations (or an analysis of failure cases), and from reporting the IRL=0 baseline on the full Val-Unseen split to show that imagination helps beyond the tuning subset.
minor comments (4)
  1. [Eq. (3)] The encoder names appear reversed: the text says RGB is encoded by f_DEP-ViT and depth by f_RGB-ViT, which is likely a typo. Please swap the names so that RGB uses the RGB ViT and depth uses the DepthAnything ViT.
  2. [Sec. IV-B] The sentence 'the OSR metrics are directly associated with TL' is not fully justified. OSR depends on whether any point along the path is within the success threshold, which is related to path length but not determined by TL alone. Please rephrase or provide a citation.
  3. [Fig. 5] The ablation curves in Fig. 5 show SR/SPL versus CTN and IRL, but the exact numerical values are not visible. Including a small table or axis labels with actual numbers would make the choice of CTN=4 and IRL=18 more transparent.
  4. [Table II] The baseline name 'Navid' should be capitalized consistently as 'NAViD' to match the reference.

Circularity Check

0 steps flagged

No circular derivation: DreamNav's pipeline uses external pretrained components and hand-designed procedures; no fitted parameter is renamed as a prediction and no self-citation chain forces the result.

full rationale

The derivation chain is a pipeline of external pretrained models and hand-designed procedures: Eq. (1)-(2) are thresholded mask gates; Eq. (3)-(5) are a pretrained diffusion policy conditioned on RGB-D; Eq. (6)-(7) are farthest-first geometric selection; Eq. (8) is a conditioned generative rollout. None of these outputs is defined in terms of the headline SR/SPL, and no parameter is fitted to the benchmark and then reported as a prediction. The only internal parameter choices, CTN=4 and IRL=18, are selected using 100 Val-Unseen episodes (Sec. IV-D, Fig. 5) and then applied to the full Val-Unseen table; this is a statistical evaluation-protocol concern (possible selection bias, no error bars), not a circular derivation, because the final success counts are not algebraically determined by those choices. The no-imagination ablation (IRL=0) in Fig. 5(b) and the EgoView Corrector ablations in Table III provide internal checks of the contribution of the novel modules. The paper's stated limitation of higher NE due to missing global egocentric context (Sec. IV-B) is a performance caveat, not a circular step. There are no load-bearing self-citations: all cited modules are external components, and no 'uniqueness theorem' is used to forbid alternatives.

Axiom & Free-Parameter Ledger

4 free parameters · 5 axioms · 0 invented entities

No new physical entities, forces, or conserved quantities are introduced. Dream Walker and Narration Expert are software modules composed from cited pretrained models. The main unverified assumptions are about the fidelity of pretrained world models and the transfer of pretrained diffusion policies to unseen VLN environments.

free parameters (4)
  • Micro-Adjust threshold theta = 0.1
    Used in Eq. (1) to gate post-action reorientation; hand-set in Sec. IV-A without a held-out sweep.
  • Candidate Trajectory Number (CTN) = 4
    Chosen by ablation on Val-Unseen episodes (Fig. 5a); it balances diversity and cost but is tuned on the evaluation split.
  • Imagination Rollout Length (IRL) = 18
    Chosen by ablation on Val-Unseen episodes (Fig. 5b); 18 maximizes SR/SPL on the same split used for final results.
  • Rotation amounts and max corrections = 30 degrees / 90 degrees, max 2/3 turns
    Hand-set constants in Micro/Macro adjust that determine how aggressively the agent reorients; no sensitivity analysis.
axioms (5)
  • domain assumption Stable Virtual Camera's pretrained world model produces coherent egocentric frames for unseen indoor trajectories when conditioned on pose sequences (Eq. 8).
    The entire Imagination Predictor and the Navigator's comparative semantics depend on the fidelity of these generated rollouts; the paper does not validate them against ground truth.
  • domain assumption NavDP diffusion policy's trajectory predictions transfer from its training environments to zero-shot VLN-CE instructions and scenes.
    The Trajectory Predictor is used without fine-tuning; if the diffusion head emits infeasible trajectories, the filter and downstream selection inherit the error.
  • domain assumption Walkable-area mask occupancy is a reliable proxy for heading misalignment (Eq. 1).
    The Micro-Adjust Controller turns 30 degrees based on comparing left/right walkability, assuming walls occlude walkable area only when orientation is wrong.
  • domain assumption LLM/VLM judges (GPT-4o, Qwen-VL) can reason reliably over egocentric images and imagined narratives with only the provided prompts.
    Navigation decisions, orientation corrections, and trajectory selection all route through these APIs; prompt sensitivity is not quantified.
  • domain assumption Farthest-first trajectory selection preserves the semantically distinct branches of the diffusion output.
    Eqs. (6)-(7) use mean per-step Euclidean distance as a diversity metric; no evidence that geometric dissimilarity corresponds to semantic or instruction-relevant distinctness.

pith-pipeline@v1.3.0-alltime-deepseek · 12242 in / 12269 out tokens · 138529 ms · 2026-08-04T16:55:06.291189+00:00 · methodology

0 comments
read the original abstract

Vision-and-Language Navigation in Continuous Environments (VLN-CE), which links language instructions to perception and control in the real world, is a core capability of embodied robots. Recently, large-scale pretrained foundation models have been leveraged as shared priors for perception, reasoning, and action, enabling zero-shot VLN without task-specific training. However, existing zero-shot VLN methods depend on costly perception and passive scene understanding, collapsing control to point-level choices. As a result, they are expensive to deploy, misaligned in action semantics, and short-sighted in planning. To address these issues, we present DreamNav that focuses on the following three aspects: (1) for reducing sensory cost, our EgoView Corrector aligns viewpoints and stabilizes egocentric perception; (2) instead of point-level actions, our Trajectory Predictor favors global trajectory-level planning to better align with instruction semantics; and (3) to enable anticipatory and long-horizon planning, we propose an Imagination Predictor to endow the agent with proactive thinking capability. On VLN-CE and real-world tests, DreamNav sets a new zero-shot state-of-the-art (SOTA), outperforming the strongest egocentric baseline with extra information by up to 7.49\% and 18.15\% in terms of SR and SPL metrics. To our knowledge, this is the first zero-shot VLN method to unify trajectory-level planning and active imagination while using only egocentric inputs.

Figures

Figures reproduced from arXiv: 2509.11197 by Peiran Liu, Renjing Xu, Shuning Zhang, Taowen Wang, Yawen Tan, Yiding Ji, Yixiao Feng, Yuetong Fang, Yunheng Wang.

Figure 1
Figure 1. Figure 1: Most of the prior approaches currently heavily rely on point-level actions [10–12]: the agent samples multiple candidate points from high-level panoramic observations, arXiv:2509.11197v1 [cs.RO] 14 Sep 2025 [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Pipeline of DreamNav. Our approach is structured into four hierarchical modules, each instantiated as a pair of mutually coupled submodules. It takes egocentric RGB-D observations and a natural language instruction as input. The EgoView Corrector first aligns the agent’s viewpoint with the instruction-indicated orientation. The Trajectory Predictor then generates candidate trajectories, and the Imagination… view at source ↗
Figure 3
Figure 3. Figure 3: Examples of two types of viewpoint errors encountered by an egocentric agent. (a) At initialization, the agent fails to perceive the instruction-indicated landmark (Initialization Misorientation Error) and is reoriented by the Macro-Adjust Expert to reveal the intended landmark. (b) After action execution, the agent’s view becomes occluded (Post-Action Misorientation Error) and is corrected by the Micro-Ad… view at source ↗
Figure 4
Figure 4. Figure 4: Representative DreamNav navigation examples in real-world settings: (a) Office, (b) Corridor, (c) Classroom, and (d) Auditorium. For each scene, the top row shows the navigation instruction, the middle row presents the egocentric observations used for execution, and the bottom row depicts the corresponding exocentric (third-person) trajectory. The middle-row views correspond directly to the numbered steps … view at source ↗
Figure 5
Figure 5. Figure 5: Ablation analysis of two key hyperparameters in DreamNav: (a) Candidate Trajectory Number and (b) Imagination Rollout Length. facilitate effective egocentric navigation. Notably, the com￾plete setting (iv) (SR: 35%; SPL: 30.05%) outperforms all other variants by at least 19% in SR and 15.88% in SPL. Effect of Candidate Trajectory Number. In our Tra￾jectory Predictor, CTN balances the quality of candidate t… view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 7 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Beyond Waypoints: A Trajectory-Centric Waypointing Paradigm for Vision-Language Navigation

    cs.RO 2026-06 unverdicted novelty 7.0

    The paper introduces a Trajectory Waypoint paradigm with a TSDF-guided diffusion policy and trajectory-enhanced navigator that achieves better performance on VLN-CE benchmarks by ensuring waypoint reachability and pla...

  2. Bridging the 2D-3D Gap: A Hierarchical Semantic-Geometric Map for Vision Language Navigation

    cs.CV 2026-05 unverdicted novelty 6.0

    HSGM structures 3D geometry and semantics into a multi-level map that lets VLMs perform high-level planning in zero-shot VLN, achieving SOTA on R2R-CE and RxR-CE.

  3. GA-VLN: Geometry-Aware BEV Representation for Efficient Vision-Language Navigation

    cs.CV 2026-05 unverdicted novelty 6.0

    GA-VLN builds a geometry-aware BEV representation from RGB-D inputs plus 3D foundation model features to deliver state-of-the-art vision-language navigation using only navigation data.

  4. Human Cognition in Machines: A Unified Perspective of World Models

    cs.RO 2026-04 unverdicted novelty 6.0

    The paper introduces a unified framework for world models that fully incorporates all cognitive functions from Cognitive Architecture Theory, highlights under-researched areas in motivation and meta-cognition, and pro...

  5. What Limits Vision-and-Language Navigation ?

    cs.RO 2026-05 unverdicted novelty 5.0

    StereoNav reaches new benchmark highs on R2R-CE and RxR-CE and improves real-robot reliability by supplying persistent target-location priors and stereo-derived geometry that stay stable under lighting changes and blur.

  6. LightZeroNav: Zero-Shot Vision Language Navigation in Continuous Environments Based on Lightweight VLMs

    cs.CV 2026-03 unverdicted novelty 5.0

    LightZeroNav decomposes zero-shot VLN-CE into modules that reduce input redundancy, improve progress tracking from noisy memory, and separate action execution from stage transitions, allowing an 8B VLM to match GPT-4o...

  7. Towards Dual-Brain Minimal Sufficient Representation for Vision-Language Navigation

    cs.CV 2026-07 reject novelty 4.0

    A CP-decomposed, instruction-conditioned latent bottleneck (CompactNav) improves VLN-CE success rate by about 2% over prior state of the art on two benchmarks.

Reference graph

Works this paper leans on

39 extracted references · 18 linked inside Pith · cited by 7 Pith papers

  1. [1]

    Beyond the nav-graph: Vision and language navigation in continuous environ- ments,

    J. Krantz, E. Wijmans, A. Majumdar, D. Batra, and S. Lee, “Beyond the nav-graph: Vision and language navigation in continuous environ- ments,” inEuropean Conference on Computer Vision (ECCV), 2020, pp. 104–120

    2020

  2. [2]

    Vision-and-language navigation: Interpreting visually-grounded navigation instructions in real environments,

    P. Anderson, Q. Wu, D. Teney, J. Bruce, M. Johnson, N. S ¨underhauf, I. Reid, S. Gould, and A. van den Hengel, “Vision-and-language navigation: Interpreting visually-grounded navigation instructions in real environments,” inIEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018, pp. 3674–3683

    2018

  3. [3]

    Vision- and-language navigation: A survey of tasks, methods, and future directions,

    J. Gu, E. Stefani, Q. Wu, J. Thomason, and X. E. Wang, “Vision- and-language navigation: A survey of tasks, methods, and future directions,” inAnnual Meeting of the Association for Computational Linguistics (ACL), 2022

    2022

  4. [4]

    Vln-r1: Vision- language navigation via reinforcement fine-tuning,

    Z. Qi, Z. Zhang, Y . Yu, J. Wang, and H. Zhao, “Vln-r1: Vision- language navigation via reinforcement fine-tuning,”arXiv preprint arXiv:2506.17221, 2025

    Pith/arXiv arXiv 2025

  5. [5]

    Navid: Video-based vlm plans the next step for vision-and-language navigation,

    J. Zhang, K. Wang, R. Xu, G. Zhou, Y . Hong, X. Fang, Q. Wu, Z. Zhang, and H. Wang, “Navid: Video-based vlm plans the next step for vision-and-language navigation,” inRobotics: Science and Systems (RSS), 2024

    2024

  6. [6]

    Uni-navid: A video-based vision-language- action model for unifying embodied navigation tasks,

    J. Zhang, K. Wang, S. Wang, M. Li, H. Liu, S. Wei, Z. Wang, Z. Zhang, and H. Wang, “Uni-navid: A video-based vision-language- action model for unifying embodied navigation tasks,” inRobotics: Science and Systems (RSS), 2024

    2024

  7. [7]

    Mp5: A multi-modal open-ended embodied system in minecraft via active perception,

    Y . Qin, E. Zhou, Q. Liu, Z. Yin, L. Sheng, R. Zhang, Y . Qiao, and J. Shao, “Mp5: A multi-modal open-ended embodied system in minecraft via active perception,” inIEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2024, pp. 16 307–16 316

    2024

  8. [8]

    Navgpt: Explicit reasoning in vision- and-language navigation with large language models,

    G. Zhou, Y . Hong, and Q. Wu, “Navgpt: Explicit reasoning in vision- and-language navigation with large language models,”arXiv preprint arXiv:2305.16986, 2023

    Pith/arXiv arXiv 2023

  9. [9]

    Mapgpt: Map-guided prompting with adaptive path planning for vision-and-language navigation,

    J. Chen, B. Lin, R. Xu, Z. Chai, X. Liang, and K.-Y . K. Wong, “Mapgpt: Map-guided prompting with adaptive path planning for vision-and-language navigation,” inAnnual Meeting of the Association for Computational Linguistics (ACL), 2024

    2024

  10. [10]

    Open-nav: Exploring zero-shot vision-and-language navigation in continuous environment with open-source llms,

    Y . Qiao, W. Lyu, H. Wang, Z. Wang, Z. Li, Y . Zhang, M. Tan, and Q. Wu, “Open-nav: Exploring zero-shot vision-and-language navigation in continuous environment with open-source llms,” inIEEE International Conference on Robotics and Automation (ICRA), 2025

    2025

  11. [11]

    Smart- way: Enhanced waypoint prediction and backtracking for zero-shot vision-and-language navigation,

    X. Shi, Z. Li, W. Lyu, J. Xia, F. Dayoub, Y . Qiao, and Q. Wu, “Smart- way: Enhanced waypoint prediction and backtracking for zero-shot vision-and-language navigation,”arXiv preprint arXiv:2503.10069, 2025

    Pith/arXiv arXiv 2025

  12. [12]

    Instructnav: Zero-shot system for generic instruction navigation in unexplored environment,

    Y . Long, W. Cai, H. Wang, G. Zhan, and H. Dong, “Instructnav: Zero-shot system for generic instruction navigation in unexplored environment,”arXiv preprint arXiv:2406.04882, 2024

    Pith/arXiv arXiv 2024

  13. [13]

    View invariant learning for vision-language navigation in continuous environments,

    J. Q. Sun, X. Xing, H. Weng, C. M. Yeum, and M. Crowley, “View invariant learning for vision-language navigation in continuous environments,”arXiv preprint arXiv:2507.08831, 2025

    arXiv 2025

  14. [14]

    Mc-gpt: Empowering vision-and- language navigation with memory map and reasoning chains,

    Z. Zhan, L. Yu, S. Yu, and G. Tan, “Mc-gpt: Empowering vision-and- language navigation with memory map and reasoning chains,”arXiv preprint arXiv:2405.10620, 2024

    Pith/arXiv arXiv 2024

  15. [15]

    Constraint-aware zero-shot vision-language navigation in continuous environments,

    K. Chen, D. An, Y . Huang, R. Xu, Y . Su, Y . Ling, I. Reid, and L. Wang, “Constraint-aware zero-shot vision-language navigation in continuous environments,”arXiv preprint arXiv:2412.10137, 2024

    Pith/arXiv arXiv 2024

  16. [16]

    Bridging the gap between learning in discrete and continuous environments for vision-and- language navigation,

    Y . Hong, Z. Wang, Q. Wu, and S. Gould, “Bridging the gap between learning in discrete and continuous environments for vision-and- language navigation,” inIEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2022, pp. 15 439–15 449

    2022

  17. [17]

    Path- dreamer: A world model for indoor navigation,

    J. Y . Koh, H. Lee, Y . Yang, J. Baldridge, and P. Anderson, “Path- dreamer: A world model for indoor navigation,” inIEEE International Conference on Computer Vision (ICCV), 2021, pp. 14 738–14 748

    2021

  18. [18]

    Dreamwalker: Mental planning for continuous vision-language navigation,

    H. Wang, W. Liang, L. V . Gool, and W. Wang, “Dreamwalker: Mental planning for continuous vision-language navigation,” inIEEE Inter- national Conference on Computer Vision (ICCV), 2023, pp. 10 873– 10 883

    2023

  19. [19]

    Navigation world models,

    A. Bar, G. Zhou, D. Tran, T. Darrell, and Y . LeCun, “Navigation world models,”arXiv preprint arXiv:2412.03572, 2024

    Pith/arXiv arXiv 2024

  20. [20]

    Imagine before go: Self-supervised generative map for object goal navigation,

    S. Zhang, X. Yu, X. Song, X. Wang, and S. Jiang, “Imagine before go: Self-supervised generative map for object goal navigation,” inIEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2024, pp. 16 414–16 425

    2024

  21. [21]

    Peanut: Predicting and navigating to unseen targets,

    A. J. Zhai and S. Wang, “Peanut: Predicting and navigating to unseen targets,” inIEEE International Conference on Computer Vision (ICCV), 2023, pp. 10 926–10 935

    2023

  22. [22]

    Unitedvln: Generalizable gaussian splatting for continuous vision-language navigation,

    G. Dai, J. Zhao, Y . Chen, Y . Qin, H. Zhao, G. Xie, Y . Yao, X. Shu, and X. Li, “Unitedvln: Generalizable gaussian splatting for continuous vision-language navigation,”arXiv preprint arXiv:2411.16053, 2024

    Pith/arXiv arXiv 2024

  23. [23]

    Looka- head exploration with neural radiance representation for continuous vision-language navigation,

    Z. Wang, X. Li, J. Yang, Y . Liu, J. Hu, M. Jiang, and S. Jiang, “Looka- head exploration with neural radiance representation for continuous vision-language navigation,” inIEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2024, pp. 13 753–13 762

    2024

  24. [24]

    Gpt-4 technical report,

    J. Achiam, S. Adler, S. Agarwal, L. Ahmad, I. Akkaya, F. L. Aleman, D. Almeida, J. Altenschmidt, S. Altman, S. Anadkat,et al., “Gpt-4 technical report,”arXiv preprint arXiv:2303.08774, 2023

    Pith/arXiv arXiv 2023

  25. [25]

    Fast segment anything,

    X. Zhao, W. Ding, Y . An, Y . Du, T. Yu, M. Li, M. Tang, and J. Wang, “Fast segment anything,”arXiv preprint arXiv:2306.12156, 2023

    Pith/arXiv arXiv 2023

  26. [26]

    Diffusion policy: Visuomotor policy learning via action diffusion,

    C. Chi, S. Feng, Y . Du, Z. Xu, E. Cousineau, B. Burchfiel, and S. Song, “Diffusion policy: Visuomotor policy learning via action diffusion,” in Robotics: Science and Systems (RSS), 2023

    2023

  27. [27]

    Diffusionvla: Scaling robot founda- tion models via unified diffusion and autoregression,

    J. Wen, M. Zhu, Y . Zhu, Z. Tang, J. Li, Z. Zhou, C. Li, X. Liu, Y . Peng, C. Shen, and F. Feng, “Diffusionvla: Scaling robot founda- tion models via unified diffusion and autoregression,”arXiv preprint arXiv:2412.03293, 2025

    Pith/arXiv arXiv 2025

  28. [28]

    Nomad: Goal masked diffusion policies for navigation and exploration,

    A. Sridhar, D. Shah, C. Glossop, and S. Levine, “Nomad: Goal masked diffusion policies for navigation and exploration,”arXiv preprint arXiv:2310.07896, 2023

    Pith/arXiv arXiv 2023

  29. [29]

    Ldp: A local diffusion planner for efficient robot navigation and collision avoidance,

    W. Yu, J. Peng, H. Yang, J. Zhang, Y . Duan, J. Ji, and Y . Zhang, “Ldp: A local diffusion planner for efficient robot navigation and collision avoidance,”arXiv preprint arXiv:2407.01950, 2024

    Pith/arXiv arXiv 2024

  30. [30]

    Navdp: Learning sim-to-real navigation dif- fusion policy with privileged information guidance,

    W. Cai, J. Peng, Y . Yang, Y . Zhang, M. Wei, H. Wang, Y . Chen, T. Wang, and J. Pang, “Navdp: Learning sim-to-real navigation dif- fusion policy with privileged information guidance,”arXiv preprint arXiv:2505.08712, 2025

    arXiv 2025

  31. [31]

    Depth anything: Unleashing the power of large-scale unlabeled data,

    L. Yang, B. Kang, Z. Huang, X. Xu, J. Feng, and H. Zhao, “Depth anything: Unleashing the power of large-scale unlabeled data,” inIEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2024

    2024

  32. [32]

    U-net: Convolu- tional networks for biomedical image segmentation,

    O. Ronneberger, P. Fischer, and T. Brox, “U-net: Convolu- tional networks for biomedical image segmentation,”arXiv preprint arXiv:1505.04597, 2015

    Pith/arXiv arXiv 2015

  33. [33]

    Denoising diffusion probabilistic models,

    J. Ho, A. Jain, and P. Abbeel, “Denoising diffusion probabilistic models,”arXiv preprint arXiv:2006.11239, 2020

    Pith/arXiv arXiv 2006

  34. [34]

    Clustering to minimize the maximum intercluster distance,

    T. F. Gonzalez, “Clustering to minimize the maximum intercluster distance,”Theoretical Computer Science, vol. 38, pp. 293–306, 1985

    1985

  35. [35]

    Stable virtual camera: Generative view synthesis with diffusion models,

    J. Zhou, H. Gao, V . V oleti, A. Vasishta, C.-H. Yao, M. Boss, P. Torr, C. Rupprecht, and V . Jampani, “Stable virtual camera: Generative view synthesis with diffusion models,”arXiv preprint arXiv:2503.14489, 2025

    Pith/arXiv arXiv 2025

  36. [36]

    Habitat: A platform for embodied ai research,

    M. Savva, A. Kadian, O. Maksymets, Y . Zhao, E. Wijmans, B. Jain, J. Straub, J. Liu, V . Koltun, J. Malik, D. Parikh, and D. Batra, “Habitat: A platform for embodied ai research,” inIEEE International Conference on Computer Vision (ICCV), 2019

    2019

  37. [37]

    Discuss before moving: Visual language navigation via multi-expert discussions,

    Y . Long, X. Li, W. Cai, and H. Dong, “Discuss before moving: Visual language navigation via multi-expert discussions,” inIEEE International Conference on Robotics and Automation (ICRA), 2023, pp. 17 380–17 387

    2023

  38. [38]

    Gpt-4o system card,

    A. Hurst, A. Lerer, A. P. Goucher, A. Perelman, A. Ramesh, A. Clark, A. Ostrow, A. Welihinda, A. Hayes, A. Radford,et al., “Gpt-4o system card,”arXiv preprint arXiv:2410.21276, 2024

    Pith/arXiv arXiv 2024

  39. [39]

    Qwen2.5-vl technical report,

    S. Bai, K. Chen, X. Liu, J. Wang, W. Ge, S. Song, K. Dang, P. Wang, S. Wang, J. Tang,et al., “Qwen2.5-vl technical report,”arXiv preprint arXiv:2502.13923, 2025

    Pith/arXiv arXiv 2025