Pith. sign in

REVIEW 3 major objections 3 minor 88 references

SkeNa: Learning to Navigate Unseen Environments Based on Abstract Hand-Drawn Maps

T0 review · 3 major / 3 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The paper claims that an embodied agent can reach goals in unseen indoor scenes using only a hand-drawn sketch map, and reports that SkeNavigator improves SPL by 105% relative over floor-plan navigation methods on the high-abstraction…

desk verdict Plausible new task and large dataset, but the headline SPL gain rests on synthetic sketches and a single unvarying number, so treat as promising but unverified. read the letter →

arxiv 2508.03053 v1 pith:SRBXS6LS submitted 2025-08-05 cs.RO cs.AI

Haojun Xu , Jiaqi Xiang , Wu Wei , Jinyu Chen , Linqing Zhong , Linjiang Huang , Hongyu Yang , Si Liu This is my paper
classification cs.ROcs.AI
keywords sketchmapnavigationembodiedgoalpredictionhand-drawnmapsfloorplandatasetSPL
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper introduces SkeNa, a navigation task in which an embodied agent must reach a goal in an indoor scene it has never visited, using only a hand-drawn sketch map as its guide. To support the task, the authors create SoR, a dataset of 54,000 trajectory–sketch pairs drawn from 71 indoor scenes, with two validation sets that differ in how faithfully the sketches preserve spatial scale. They propose SkeNavigator, a framework that predicts navigation targets by aligning sketch-map features with an exploration map built from the agent's own observations. The paper reports that SkeNavigator beats prior floor-plan navigation methods by a large margin, improving success weighted by path length (SPL) by 105% relative on the high-abstraction validation set. If the result holds, it means a rough human-style sketch can substitute for a precise floor plan in goal-directed embodied navigation.

What carries the argument

The Ray-based Map Descriptor (RMD) is the representational workhorse: it samples evenly spaced points on the sketch map and, from each point, measures distances to map boundaries along multiple directions, turning a distorted sketch into a feature vector that stays comparable across abstraction levels. The Dual-Map Aligned Goal Predictor (DAGP) is the decision mechanism: it matches RMD features of the sketch against an exploration map accumulated from the agent's observations, and from the matched correspondence predicts the goal location that the agent should move toward. Together they replace the assumption that the agent must reconstruct an accurate metric floor plan from the sketch.

What would settle it

Collect real human-drawn route sketches for the 71 SoR scenes, run SkeNavigator and the floor-plan baselines on them, and compare SPL; if the 105% relative advantage shrinks or disappears on human sketches, the synthetic sketch distribution, rather than the alignment method, is responsible for the result.

Watch

Extended reading notes

Core claim

The paper's central claim is that abstract hand-drawn route maps are sufficient guidance for embodied navigation in unseen environments, provided the agent learns to align them with the world it actually observes. SkeNavigator treats the problem as goal prediction rather than map matching: the Ray-based Map Descriptor (RMD) encodes the sketch by placing equidistant sampling points and recording boundary distances along rays, which stabilizes the representation against the geometric distortion typical of sketches; the Dual-Map Aligned Goal Predictor (DAGP) then finds correspondences between those sketch features and features of an exploration map constructed online, and predicts the goal position that steers navigation. On the SoR high-abstraction validation set, this sketch-to-exploration-map alignment yields a 105% relative improvement in SPL over prior floor-plan-based methods, which is the paper's headline quantitative claim.

Load-bearing premise

The evaluation assumes that sketches generated automatically from floor plans resemble the sketches people actually draw when giving directions, with similar distortions and omissions; if real hand-drawn sketches are systematically messier, the reported gains may not transfer.

Editorial extensions

If this is right

  • A sketch map can serve as the sole navigation instruction for an agent entering an unseen environment, removing the need for metric floor plans or global positioning.
  • SoR provides 54,000 trajectory–sketch pairs across 71 scenes, giving a training and evaluation base for sketch-guided navigation research.
  • The two abstraction-level validation sets make it possible to measure how navigation quality degrades as sketches lose spatial-scale fidelity.
  • The reported 105% relative SPL improvement indicates that aligning sketches with on-site exploration maps, rather than matching them to floor-plan geometry, is the step that matters most.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The paper's sketches are generated from floor plans by an automated pipeline; a natural next test is whether genuinely human-drawn route sketches, with erratic distortion, missing landmarks, and personal style, preserve the 105% SPL advantage or erode it.
  • The RMD+DAGP design suggests a general recipe: represent any coarse directional prior as ray-based boundary distances, then align it with an incrementally built exploration map; verbal route instructions or stylized diagrammatic hints could be handled the same way.
  • If the abstraction-level splits behave monotonically, an explicit degradation curve would reveal whether the method has a floor below which sketch guidance stops helping.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 3 minor

Summary. The paper introduces SkeNa, a new embodied navigation task in which an agent must use a hand-drawn sketch map as the only navigational guide to reach a goal in an unseen indoor environment. The authors contribute SoR, a 54k-pair dataset of trajectories and sketch maps across 71 scenes, built by an automated pipeline that converts floor plans into hand-drawn-like sketches; the dataset includes two validation splits distinguished by the degree to which spatial scale is preserved. On top of this, they propose SkeNavigator, which uses a Ray-based Map Descriptor (RMD) to encode sketch features and a Dual-Map Aligned Goal Predictor (DAGP) to align sketch features with an exploration map and predict the goal. The paper reports that SkeNavigator outperforms prior floor-plan navigation methods, with a 105% relative SPL improvement on the high-abstraction validation set. The full text provided for review is heavily garbled, so the detailed method and experimental protocol could not be fully inspected; this report is based primarily on the abstract and the readable fragments.

Significance. If the reported result is robust, the paper makes a useful contribution: it introduces a new task, a large dataset, and a concrete method, and it commits to releasing code and data. The idea of aligning abstract sketches with online exploration maps is intuitive and potentially valuable for human-robot communication. However, the significance as stated depends on two things the current version does not establish: (i) that the automated sketches are representative of real human hand-drawn maps, and (ii) that the 105% relative improvement is statistically meaningful. The paper could become an important benchmark if those gaps are addressed.

major comments (3)
  1. [SoR dataset (construction)] The central claim of a 105% relative SPL improvement is measured entirely on the SoR validation sets, which are generated by the authors' automated floor-plan-to-sketch pipeline. Because real hand-drawn route maps typically distort spatial scale, rotate local geometry, and omit landmarks, while the synthetic pipeline preserves metric structure, the abstraction-level validation sets may not reflect the difficulty of human sketches. Since the proposed RMD explicitly uses equidistant sampling points and boundary distances, it is tuned to exactly the cues that metric-preserving synthetic sketches provide. The paper needs either (a) an evaluation on human-collected sketches, or (b) a strong argument and analysis demonstrating that the synthetic sketches approximate the distribution of human sketches; otherwise the headline result remains specific to the synthetic benchmark, not the motivating scenario.
  2. [Results / Evaluation] The reported SPL improvement is a single number with no error bars, no number of independent runs or seeds, and no significance test. If the agent's policy is stochastic or the environment initialization varies, the paper should report means and standard deviations over multiple seeds and perform a paired significance test; if the evaluation is deterministic, the paper should state this explicitly. Without such reporting, the 105% relative improvement cannot be distinguished from noise.
  3. [Baselines / Experimental setup] The abstract states that SkeNavigator outperforms 'prior floor plan navigation methods,' but it is unclear whether those baselines were trained and evaluated on the same sketch input or on floor-plan inputs adapted to the new task. A fair comparison against prior methods requires either that the baselines consume the same sketch maps (with suitable modifications) or that the paper clearly labels the comparison as a separate upper-bound analysis. The details necessary to check this are not legible in the available text; the authors should clarify the comparison protocol and report baseline performance on the same validation splits.
minor comments (3)
  1. [Abstract] The abstract reports 'improving SPL ... by 105% relatively' without giving the absolute SPL values for the new method and the baselines; please include absolute numbers on both abstraction-level validation sets.
  2. [Dataset description] The two abstraction-level validation sets are categorized 'based on their preservation of spatial scales' but the thresholds and the exact procedure are not described in any text I could read; please define the categorization rule precisely.
  3. [Code/Dataset release] The paper states that code and data 'will be released' but provides no URL or anonymized access mechanism; for review reproducibility, please include a link in the submission or an anonymized appendix describing the release plan.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity; the self-built SoR benchmark is a generalizability caveat, not a tautology.

full rationale

The central claim of arXiv:2508.03053 is empirical: SkeNavigator outperforms prior floor-plan navigation methods on the SoR validation sets. This is not a derivation from a definition. The SoR dataset is constructed by the authors' automated floor-plan-to-sketch pipeline, and the abstraction-level splits are author-defined, but that affects external validity and generalizability, not circularity. Nothing in the provided text indicates that RMD or DAGP is fitted to the validation outcomes in a way that makes the reported SPL improvement true by construction. The RMD and DAGP components are described as architectural choices; no equation in the supplied text reduces the predicted goal to the input sketch by construction. The comparison against prior floor-plan navigation methods provides independent empirical grounding. The absence of human-drawn validation sketches is a legitimate data-distribution concern, but under the circularity rubric it is a correctness or external-validity issue, not a self-referential derivation. No self-citation chain, imported uniqueness theorem, or fitted-input-called-prediction pattern is present. Score 1 reflects the minor caveat that the evaluation benchmark is entirely self-constructed, while the claim itself retains independent content.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

Review is based on the abstract only; the full text is garbled in the provided extraction. The ledger records the design choices visible from the abstract. The task definition is a domain assumption. Training and evaluation both come from the authors' own SoR dataset built by their automated pipeline, so the dataset's representativeness of human sketches is an unverified premise. No numeric free parameters are stated in the abstract; the sketch abstraction thresholds and RMD sampling count are the visible hand-chosen design choices. No invented entities beyond task artifacts (the SoR dataset and abstraction-level sets) are introduced.

free parameters (2)
  • Sketch abstraction-level categorization thresholds = not stated
    The two validation sets are separated by how much they preserve spatial scale; the thresholds for this split are author-defined and not quantified in the abstract.
  • Ray-based Map Descriptor sampling configuration = not stated
    RMD uses equidistant sampling points and boundary distances; the number of rays or sampling points is a hyperparameter that the abstract does not specify.
assumptions (3)
  • domain assumption A hand-drawn sketch map contains sufficient spatial layout information for goal-directed navigation in an unseen environment.
    This is the premise of the new SkeNa task; it defines the problem as solvable in principle. Invoked by the task definition in the abstract.
  • ad hoc to paper Automatically generated sketch representations are representative of human hand-drawn route sketches.
    SoR is built by an automated floor-plan-to-sketch pipeline; the claim that the method works for hand-drawn maps depends on this transfer, and the abstract does not state that human sketches appear in the training set.
  • ad hoc to paper The two abstraction-level validation sets, categorized by spatial-scale preservation, produce a meaningful difficulty ordering for evaluating navigation.
    The categorization thresholds are author-defined and their validity as a difficulty measure is asserted rather than derived.

how reviews work

0 comments
Cite this review

Pith. "Pith review of SkeNa: Learning to Navigate Unseen Environments Based on Abstract Hand-Drawn Maps." pith.science (2026). https://pith.science/paper/SRBXS6LS

@misc{pith2026250803053,
  author       = {Pith},
  title        = {Pith review of: SkeNa: Learning to Navigate Unseen Environments Based on Abstract Hand-Drawn Maps},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SRBXS6LS}},
  note         = {Machine review of arXiv:2508.03053}
}
read the original abstract

A typical human strategy for giving navigation guidance is to sketch route maps based on the environmental layout. Inspired by this, we introduce Sketch map-based visual Navigation (SkeNa), an embodied navigation task in which an agent must reach a goal in an unseen environment using only a hand-drawn sketch map as guidance. To support research for SkeNa, we present a large-scale dataset named SoR, comprising 54k trajectory and sketch map pairs across 71 indoor scenes. In SoR, we introduce two navigation validation sets with varying levels of abstraction in hand-drawn sketches, categorized based on their preservation of spatial scales in the environment, to facilitate future research. To construct SoR, we develop an automated sketch-generation pipeline that efficiently converts floor plans into hand-drawn representations. To solve SkeNa, we propose SkeNavigator, a navigation framework that aligns visual observations with hand-drawn maps to estimate navigation targets. It employs a Ray-based Map Descriptor (RMD) to enhance sketch map valid feature representation using equidistant sampling points and boundary distances. To improve alignment with visual observations, a Dual-Map Aligned Goal Predictor (DAGP) leverages the correspondence between sketch map features and on-site constructed exploration map features to predict goal position and guide navigation. SkeNavigator outperforms prior floor plan navigation methods by a large margin, improving SPL on the high-abstract validation set by 105% relatively. Our code and dataset will be released.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

88 extracted references · 38 canonical work pages

  1. [1]

    Anderson, P.; Wu, Q.; Teney, D.; Bruce, J.; Johnson, M.; S \"u nderhauf, N.; Reid, I.; Gould, S.; and Van Den Hengel, A. 2018. Vision-and-language navigation: Interpreting visually-grounded navigation instructions in real environments. In Proceedings of the IEEE conference on computer vision and pattern recognition, 3674--3683

    2018

  2. [2]

    Boniardi, F.; Caselitz, T.; K \"u mmerle, R.; and Burgard, W. 2019 a . A pose graph-based localization system for long-term navigation in CAD floor plans. Robotics and Autonomous Systems, 112: 84--97

    2019

  3. [3]

    Boniardi, F.; Valada, A.; Mohan, R.; Caselitz, T.; and Burgard, W. 2019 b . Robot localization in floor plans using a room layout edge extraction network. In 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 5291--5297. IEEE

    2019

  4. [5]

    S.; Gandhi, D

    Chaplot, D. S.; Gandhi, D. P.; Gupta, A.; and Salakhutdinov, R. R. 2020 a . Object goal navigation using goal-oriented semantic exploration. Advances in Neural Information Processing Systems, 33: 4247--4258

    2020

  5. [6]

    S.; Salakhutdinov, R.; Gupta, A.; and Gupta, S

    Chaplot, D. S.; Salakhutdinov, R.; Gupta, A.; and Gupta, S. 2020 b . Neural topological slam for visual navigation. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 12875--12884

    2020

  6. [7]

    Chen, C.; Wang, R.; Vogel, C.; and Pollefeys, M. 2025. F\ 3\ Loc : Fusion and Filtering for Floorplan Localization . arXiv:2403.03370

    arXiv 2025

  7. [8]

    Chen, J.; Wang, W.; Liu, S.; Li, H.; and Yang, Y. 2023. Omnidirectional Information Gathering for Knowledge Transfer-based Audio-Visual Navigation . In 2023 IEEE / CVF International Conference on Computer Vision ( ICCV ) , 10959--10969. Paris, France: IEEE. ISBN 979-8-3503-0718-4

    2023

  8. [9]

    K.; and Chen, T

    Chu, H.; Kim, D. K.; and Chen, T. 2015. You are here: Mimicking the human thinking process in reading floor-plans. In Proceedings of the IEEE International Conference on Computer Vision, 2210--2218

    2015

Show all 88 references
  1. [11]

    Elfes, A. 1989. Using occupancy grids for mobile robot perception and navigation. Computer, 22(6): 46--57

    1989

  2. [12]

    Y.; Wortsman, M.; Ilharco, G.; Schmidt, L.; and Song, S

    Gadre, S. Y.; Wortsman, M.; Ilharco, G.; Schmidt, L.; and Song, S. 2023. Cows on pasture: Baselines and benchmarks for language-driven zero-shot object navigation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 23171--23181

    2023

  3. [13]

    S.; Tulsiani, S.; Mukadam, M.; Rehg, J

    Hahn, M.; Chaplot, D. S.; Tulsiani, S.; Mukadam, M.; Rehg, J. M.; and Gupta, A. 2021. No rl, no simulation: Learning to navigate without navigating. Advances in Neural Information Processing Systems, 34: 26661--26673

    2021

  4. [14]

    Huang, C.; Mees, O.; Zeng, A.; and Burgard, W. 2023. Visual language maps for robot navigation. In 2023 IEEE International Conference on Robotics and Automation (ICRA), 10608--10615. IEEE

    2023

  5. [15]

    D.; Stachniss, C.; and Burgard, W

    Ito, S.; Endres, F.; Kuderer, M.; Tipaldi, G. D.; Stachniss, C.; and Burgard, W. 2014. W-rgb-d: floor-plan-based indoor global localization using a depth camera and wifi. In 2014 IEEE international conference on robotics and automation (ICRA), 417--422. IEEE

    2014

  6. [16]

    Kim, N.; Kwon, O.; Yoo, H.; Choi, Y.; Park, J.; and Oh, S. 2023. Topological semantic graph memory for image-goal navigation. In Conference on Robot Learning, 393--402. PMLR

    2023

  7. [17]

    Krantz, J.; Gervet, T.; Yadav, K.; Wang, A.; Paxton, C.; Mottaghi, R.; Batra, D.; Malik, J.; Lee, S.; and Chaplot, D. S. 2023. Navigating to objects specified by images. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 10916--10925

    2023

  8. [19]

    Kwon, O.; Park, J.; and Oh, S. 2023. Renderable neural radiance map for visual navigation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 9099--9108

    2023

  9. [20]

    L.; Le Chan, J.; Li, Z.; Wan, K

    Li, J.; Chan, C. L.; Le Chan, J.; Li, Z.; Wan, K. W.; and Yau, W. Y. 2021. Cognitive navigation for indoor environment using floorplan. In 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 9030--9037. IEEE

    2021

  10. [22]

    H.; and Rus, D

    Li, Z.; Ang, M. H.; and Rus, D. 2020. Online localization with imprecise floor space maps using stochastic gradient descent. In 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 8571--8578. IEEE

    2020

  11. [23]

    Mendez, O.; Hadfield, S.; Pugeault, N.; and Bowden, R. 2020. Sedar: Reading floorplans like a human—using deep learning to enable human-inspired localisation. International Journal of Computer Vision, 128(5): 1286--1310

    2020

  12. [24]

    Partsey, R.; Wijmans, E.; Yokoyama, N.; Dobosevych, O.; Batra, D.; and Maksymets, O. 2022. Is mapping necessary for realistic pointgoal navigation? In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 17232--17241

    2022

  13. [25]

    K.; Chaplot, D

    Ramakrishnan, S. K.; Chaplot, D. S.; Al-Halah, Z.; Malik, J.; and Grauman, K. 2022. Poni: Potential functions for objectgoal navigation with interaction-free learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 18890--18900

    2022

  14. [27]

    Savva, M.; Kadian, A.; Maksymets, O.; Zhao, Y.; Wijmans, E.; Jain, B.; Straub, J.; Liu, J.; Koltun, V.; Malik, J.; et al. 2019. Habitat: A platform for embodied ai research. In Proceedings of the IEEE/CVF international conference on computer vision, 9339--9347

    2019

  15. [29]

    Setalaphruk, V.; Ueno, A.; Kume, I.; Kono, Y.; and Kidode, M. 2003. Robot navigation in corridor environments using a sketch floor map. In Proceedings 2003 IEEE International Symposium on Computational Intelligence in Robotics and Automation. Computational Intelligence in Robo...

    2003

  16. [30]

    H.; Wu, J.; Washington, C.; Sadler, B

    Song, C. H.; Wu, J.; Washington, C.; Sadler, B. M.; Chao, W.-L.; and Su, Y. 2023. Llm-planner: Few-shot grounded planning for embodied agents with large language models. In Proceedings of the IEEE/CVF international conference on computer vision, 2998--3009

    2023

  17. [31]

    Y.; Bachmann, R

    Vinker, Y.; Pajouheshgar, E.; Bo, J. Y.; Bachmann, R. C.; Bermano, A. H.; Cohen-Or, D.; Zamir, A.; and Shamir, A. 2022. Clipasso: Semantically-aware object sketching. ACM Transactions on Graphics (TOG), 41(4): 1--11

    2022

  18. [32]

    Wang, H.; Liang, W.; Van Gool, L.; and Wang, W. 2023. Dreamwalker: Mental planning for continuous vision-language navigation. In Proceedings of the IEEE/CVF international conference on computer vision, 10873--10883

    2023

  19. [33]

    J.; and Olson, E

    Wang, X.; Marcotte, R. J.; and Olson, E. 2019. GLFP: Global localization from a floor plan. In 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 1627--1632. IEEE

    2019

  20. [36]

    Yu, B.; Kasaei, H.; and Cao, M. 2023. L3mvn: Leveraging large language models for visual target navigation. In 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 3554--3560. IEEE

    2023

  21. [37]

    Zhou, K.; Zheng, K.; Pryor, C.; Shen, Y.; Jin, H.; Getoor, L.; and Wang, X. E. 2023. Esc: Exploration with soft commonsense constraints for zero-shot object navigation. In International Conference on Machine Learning, 42829--42842. PMLR

    2023

  22. [38]

    Zhu, F.; Liang, X.; Zhu, Y.; Yu, Q.; Chang, X.; and Liang, X. 2021. Soon: Scenario oriented object navigation with graph-based exploration. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 12689--12699

    2021

  23. [39]

    Communication, Simulation, and Intelligent Agents: Implications of Personal Intelligent Machines for Medical Education

    Clancey, William J. Communication, Simulation, and Intelligent Agents: Implications of Personal Intelligent Machines for Medical Education. Proceedings of the Eighth International Joint Conference on Artificial Intelligence (IJCAI-83)

  24. [40]

    Classification Problem Solving

    Clancey, William J. Classification Problem Solving. Proceedings of the Fourth National Conference on Artificial Intelligence

  25. [41]

    , title =

    Robinson, Arthur L. , title =. 1980 , doi =. https://science.sciencemag.org/content/208/4447/1019.full.pdf , journal =

    1980

  26. [42]

    New Ways to Make Microcircuits Smaller---Duplicate Entry

    Robinson, Arthur L. New Ways to Make Microcircuits Smaller---Duplicate Entry. Science

  27. [43]

    Clancey and Glenn Rennels , abstract =

    Diane Warner Hasling and William J. Clancey and Glenn Rennels , abstract =. Strategic explanations for a diagnostic consultation system , journal =. 1984 , issn =. doi:https://doi.org/10.1016/S0020-7373(84)80003-6 , url =

    1984 doi

  28. [44]

    and Rennels, Glenn R

    Hasling, Diane Warner and Clancey, William J. and Rennels, Glenn R. and Test, Thomas. Strategic Explanations in Consultation---Duplicate. The International Journal of Man-Machine Studies

  29. [45]

    Poligon: A System for Parallel Problem Solving

    Rice, James. Poligon: A System for Parallel Problem Solving

  30. [46]

    Transfer of Rule-Based Expertise through a Tutorial Dialogue

    Clancey, William J. Transfer of Rule-Based Expertise through a Tutorial Dialogue

  31. [47]

    The Engineering of Qualitative Models

    Clancey, William J. The Engineering of Qualitative Models

  32. [48]

    2017 , eprint=

    Attention Is All You Need , author=. 2017 , eprint=

    2017

  33. [49]

    Pluto: The 'Other' Red Planet

    NASA. Pluto: The 'Other' Red Planet

  34. [50]

    arXiv preprint arXiv:2010.07954 , year=

    Room-across-room: Multilingual vision-and-language navigation with dense spatiotemporal grounding , author=. arXiv preprint arXiv:2010.07954 , year=

    2010 arXiv

  35. [51]

    arXiv preprint arXiv:1911.00357 , year=

    Dd-ppo: Learning near-perfect pointgoal navigators from 2.5 billion frames , author=. arXiv preprint arXiv:1911.00357 , year=

    1911 arXiv

  36. [52]

    2017 IEEE international conference on robotics and automation (ICRA) , pages=

    Target-driven visual navigation in indoor scenes using deep reinforcement learning , author=. 2017 IEEE international conference on robotics and automation (ICRA) , pages=. 2017 , organization=

    2017

  37. [53]

    arXiv preprint arXiv:2109.08238 , year=

    Habitat-matterport 3d dataset (hm3d): 1000 large-scale 3d environments for embodied ai , author=. arXiv preprint arXiv:2109.08238 , year=

    arXiv

  38. [54]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Is mapping necessary for realistic pointgoal navigation? , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  39. [55]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Navigating to objects specified by images , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  40. [56]

    arXiv preprint arXiv:2303.07798 , year=

    Ovrl-v2: A simple state-of-art baseline for imagenav and objectnav , author=. arXiv preprint arXiv:2303.07798 , year=

    arXiv

  41. [57]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Neural topological slam for visual navigation , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  42. [58]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Poni: Potential functions for objectgoal navigation with interaction-free learning , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  43. [59]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Cows on pasture: Baselines and benchmarks for language-driven zero-shot object navigation , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  44. [60]

    Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

    Vision-and-language navigation: Interpreting visually-grounded navigation instructions in real environments , author=. Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

  45. [61]

    Proceedings of the IEEE/CVF international conference on computer vision , pages=

    Llm-planner: Few-shot grounded planning for embodied agents with large language models , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=

  46. [62]

    Proceedings of the IEEE/CVF international conference on computer vision , pages=

    Dreamwalker: Mental planning for continuous vision-language navigation , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=

  47. [63]

    Robotics and Autonomous Systems , volume=

    A pose graph-based localization system for long-term navigation in CAD floor plans , author=. Robotics and Autonomous Systems , volume=. 2019 , publisher=

    2019

  48. [64]

    2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , pages=

    Online localization with imprecise floor space maps using stochastic gradient descent , author=. 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , pages=. 2020 , organization=

    2020

  49. [65]

    International Journal of Computer Vision , volume=

    Sedar: Reading floorplans like a human—using deep learning to enable human-inspired localisation , author=. International Journal of Computer Vision , volume=. 2020 , publisher=

    2020

  50. [66]

    2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , pages=

    GLFP: Global localization from a floor plan , author=. 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , pages=. 2019 , organization=

    2019

  51. [67]

    2014 IEEE international conference on robotics and automation (ICRA) , pages=

    W-rgb-d: floor-plan-based indoor global localization using a depth camera and wifi , author=. 2014 IEEE international conference on robotics and automation (ICRA) , pages=. 2014 , organization=

    2014

  52. [68]

    2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , pages=

    Robot localization in floor plans using a room layout edge extraction network , author=. 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , pages=. 2019 , organization=

    2019

  53. [69]

    Proceedings of the IEEE International Conference on Computer Vision , pages=

    You are here: Mimicking the human thinking process in reading floor-plans , author=. Proceedings of the IEEE International Conference on Computer Vision , pages=

  54. [70]

    Proceedings 2003 IEEE International Symposium on Computational Intelligence in Robotics and Automation

    Robot navigation in corridor environments using a sketch floor map , author=. Proceedings 2003 IEEE International Symposium on Computational Intelligence in Robotics and Automation. Computational Intelligence in Robotics and Automation for the New Millennium (Cat. No. 03EX694)...

    2003

  55. [71]

    2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , pages=

    Cognitive navigation for indoor environment using floorplan , author=. 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , pages=. 2021 , organization=

    2021

  56. [72]

    Computer , volume=

    Using occupancy grids for mobile robot perception and navigation , author=. Computer , volume=. 1989 , publisher=

    1989

  57. [73]

    Advances in Neural Information Processing Systems , volume=

    Object goal navigation using goal-oriented semantic exploration , author=. Advances in Neural Information Processing Systems , volume=

  58. [74]

    2023 IEEE International Conference on Robotics and Automation (ICRA) , pages=

    Visual language maps for robot navigation , author=. 2023 IEEE International Conference on Robotics and Automation (ICRA) , pages=. 2023 , organization=

    2023

  59. [75]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Renderable neural radiance map for visual navigation , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  60. [76]

    Advances in Neural Information Processing Systems , volume=

    No rl, no simulation: Learning to navigate without navigating , author=. Advances in Neural Information Processing Systems , volume=

  61. [77]

    Conference on Robot Learning , pages=

    Topological semantic graph memory for image-goal navigation , author=. Conference on Robot Learning , pages=. 2023 , organization=

    2023

  62. [78]

    arXiv preprint arXiv:2412.18335 , year=

    FloNa: Floor Plan Guided Embodied Visual Navigation , author=. arXiv preprint arXiv:2412.18335 , year=

    arXiv

  63. [79]

    International Conference on Machine Learning , pages=

    Esc: Exploration with soft commonsense constraints for zero-shot object navigation , author=. International Conference on Machine Learning , pages=. 2023 , organization=

    2023

  64. [80]

    2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , pages=

    L3mvn: Leveraging large language models for visual target navigation , author=. 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , pages=. 2023 , organization=

    2023

  65. [81]

    arXiv preprint arXiv:2411.16425 , year=

    TopV-Nav: Unlocking the Top-View Spatial Reasoning Potential of MLLM for Zero-shot Object Navigation , author=. arXiv preprint arXiv:2411.16425 , year=

    arXiv

  66. [82]

    Proceedings of the IEEE/CVF international conference on computer vision , pages=

    Habitat: A platform for embodied ai research , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=

  67. [83]

    arXiv preprint arXiv:1709.06158 , year=

    Matterport3d: Learning from rgb-d data in indoor environments , author=. arXiv preprint arXiv:1709.06158 , year=

    arXiv

  68. [84]

    arXiv preprint arXiv:2412.06465 , year=

    Agent Journey Beyond RGB: Unveiling Hybrid Semantic-Spatial Environmental Representations for Vision-and-Language Navigation , author=. arXiv preprint arXiv:2412.06465 , year=

  69. [85]

    arXiv preprint arXiv:1606.02396 , year=

    Deep successor reinforcement learning , author=. arXiv preprint arXiv:1606.02396 , year=

    arXiv

  70. [86]

    arXiv preprint arXiv:1610.01733 , year=

    Towards cognitive exploration through deep reinforcement learning for mobile robots , author=. arXiv preprint arXiv:1610.01733 , year=

    arXiv

  71. [87]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Soon: Scenario oriented object navigation with graph-based exploration , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  72. [88]

    Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part VI 16 , pages=

    Soundspaces: Audio-visual navigation in 3d environments , author=. Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part VI 16 , pages=. 2020 , organization=

    2020

  73. [89]

    arXiv preprint arXiv:2008.09622 , year=

    Learning to set waypoints for audio-visual navigation , author=. arXiv preprint arXiv:2008.09622 , year=

    2008 arXiv

  74. [90]

    ACM Transactions on Graphics (TOG) , volume=

    Clipasso: Semantically-aware object sketching , author=. ACM Transactions on Graphics (TOG) , volume=. 2022 , publisher=

    2022

  75. [91]

    Proceedings of the 40th International Conference on Machine Learning , articleno =

    Zhou, Kaiwen and Zheng, Kaizhi and Pryor, Connor and Shen, Yilin and Jin, Hongxia and Getoor, Lise and Wang, Xin Eric , title =. Proceedings of the 40th International Conference on Machine Learning , articleno =. 2023 , publisher =

    2023

  76. [92]

    Chen, Changan and Wang, Rui and Vogel, Christoph and Pollefeys, Marc , year =. F\. doi:10.48550/arXiv.2403.03370 , urldate =. arXiv , langid =:2403.03370 , primaryclass =

    doi arXiv

  77. [93]

    Chung, Junyoung and Kastner, Kyle and Dinh, Laurent and Goel, Kratarth and Courville, Aaron C and Bengio, Yoshua , year =. A. Advances in

  78. [94]

    arXiv preprint arXiv:1707.06347 , year=

    Proximal policy optimization algorithms , author=. arXiv preprint arXiv:1707.06347 , year=

    arXiv

  79. [95]

    arXiv preprint arXiv:2010.11929 , year=

    An image is worth 16x16 words: Transformers for image recognition at scale , author=. arXiv preprint arXiv:2010.11929 , year=

    2010 arXiv

  80. [96]

    Omnidirectional

    Chen, Jinyu and Wang, Wenguan and Liu, Si and Li, Hongsheng and Yang, Yi , year =. Omnidirectional. 2023. doi:10.1109/ICCV51070.2023.01009 , urldate =

    2023

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.