Pith. sign in

REVIEW 4 major objections 4 minor 145 references

HoloGeo: Mitigating Landmark Bias in Geo-localization via Evidence-Driven Reasoning

T0 review · 4 major / 4 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read The paper claims landmark bias in geo-localization is measurable (Bias Intensity, Bias Harmfulness) and reducible: HoloGeo reaches 27.27% city-level accuracy on a new misleading-landmark benchmark while preserving standard performance.

desk verdict The BI/BH diagnostics and LandmarkBias-3K are genuinely new, but the BF-30K annotation prompt leaks the ground-truth location, undercutting the paper's evidence-driven claim. read the letter →

arxiv 2607.15255 v1 pith:ATHHNWOH submitted 2026-07-16 cs.CV

Pengcheng Zhou , Xuanyu Liu , Yanchen Yin , Bobo Li , Shengqiong Wu , Mong-Li Lee , Wynne Hsu This is my paper
classification cs.CV
keywords imagegeo-localizationlandmarkbiasintensityharmfulnessvision-languagemodelsmulti-evidencereasoningreinforcementlearninggeospatial
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

The paper tries to establish that a major failure mode of vision-language geo-localization—over-reliance on a single salient landmark—can be measured, benchmarked, and reduced. It introduces Bias Intensity and Bias Harmfulness, two probability-based metrics that compare model confidence on raw, landmark-removed, and landmark-only views, and uses them to build LandmarkBias-3K, a 3,000-image test set of misleading landmark scenarios. To fix the bias, the paper proposes HoloGeo, a training framework whose supervised fine-tuning and reinforcement learning rewards encourage balanced coverage of many visual cues rather than one landmark. On LandmarkBias-3K, HoloGeo reaches 27.27% city-level accuracy versus 16.83% for the 7B backbone and 23.57% for the strongest prior domain-specific model, while keeping competitive results on standard benchmarks. If right, the paper provides a reusable way to audit and correct shortcut reliance in geospatial reasoning.

What carries the argument

The load-bearing machinery is the two-metric diagnostic BI/BH, computed from three views: the raw image, the landmark-removed image, and the landmark-only crop. BI is the log-probability difference for the landmark-induced label between raw and landmark-removed inputs; BH is the change, caused by the landmark, in the log-odds of the true label versus the landmark-induced label. These metrics choose which images enter LandmarkBias-3K and which BF-30K samples are labeled landmark-biased. HoloGeo's training machinery then combines supervised fine-tuning with group-relative policy optimization, using a total reward of format compliance, hierarchical geo-accuracy (country first, city second), bou

What would settle it

Inspect the actual BF-30K annotation prompts and logs. If the confirmed country/city appears anywhere in the model input before reasoning—as the template in Figure 9 suggests—the chains are post-hoc. A decisive experiment: re-annotate a random subset without any location label, retrain HoloGeo on that subset only, and measure LandmarkBias-3K city accuracy; if it falls back toward the SFT-only level, the evidence-driven premise is refuted.

Watch

Extended reading notes

Core claim

The central claim is that landmark bias is quantifiable: Bias Intensity measures the log-confidence boost a landmark gives its favored label, and Bias Harmfulness measures how much that boost costs the ground-truth label, computed from full, landmark-removed, and landmark-only views. These metrics select LandmarkBias-3K and guide BF-30K, a training set of structured multi-evidence reasoning chains. HoloGeo fine-tunes a 7B vision-language backbone with supervised learning plus policy-gradient reinforcement learning, rewarding hierarchical geo-accuracy, visual-box coverage, and judge-scored reasoning validity. The paper reports 27.27% city-level accuracy on LandmarkBias-3K and strong results o

Load-bearing premise

The evidence-driven foundation of HoloGeo rests on the claim that the BF-30K reasoning chains were produced from visual evidence alone, without seeing the confirmed location; the annotation prompt shown in Figure 9 appears to include that confirmed location, so if that prompt was used, the chains are post-hoc rationalizations and the central debiasing premise loses its support.

Editorial extensions

If this is right

  • LandmarkBias-3K provides a diagnostic where most current geo-localization systems score below 20% city-level accuracy, exposing failures that accuracy-only benchmarks miss.
  • HoloGeo's 27.27% city-level accuracy on LandmarkBias-3K, alongside 47.3% city-level on IM2GPS and leading results on YFCC4k, indicates that reducing landmark bias does not have to sacrifice standard performance.
  • The BI/BH metric pair separates harmful anchoring from beneficial landmark use, meaning future benchmarks can be built to penalize only the harmful cases.
  • Ablations show each of the three rewards contributes; removing any one lowers performance, so the multi-reward design, not any single term, carries the improvement.
  • Cross-model tercile agreement for BI/BH suggests the metrics capture stable patterns rather than quirks of one vision-language model.

Reading between the lines

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

  • Editorial inference: The apparent leak of the confirmed location into the BF-30K annotation prompt does not by itself disprove the debiasing effect; the box-coverage and reasoning-validity rewards could be doing the work independent of chain purity. A clean test is to train on re-annotated chains without any location label and compare.
  • Editorial inference: The BI/BH construction is task-agnostic—replace 'landmark' with any dominant spurious cue and the three-view probability comparison can audit shortcut reliance in other VLM tasks, such as medical imaging or wildlife monitoring.
  • Editorial inference: The benchmark's per-category results suggest that cues like language signs and architecture are both common and misleading at city level, while climate cues are milder; balancing training data toward informative-but-not-anchor-like cue types could push accuracy further.
  • Editorial inference: The paper's own limitation section concedes that the training distribution leaves remote and underrepresented regions poorly served; this is a separate failure mode from landmark bias, so HoloGeo's robustness claim should be read as applying to well-represented regions unless the data distribution is expanded.
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

4 major / 4 minor

Summary. The paper proposes HoloGeo, a framework for mitigating landmark bias in VLM-based image geo-localization. It introduces two metrics, Bias Intensity (BI) and Bias Harmfulness (BH), and constructs a benchmark, LandmarkBias-3K, from MP-16 and GLDv2. To reduce bias, it builds a training dataset BF-30K with structured multi-evidence reasoning chains and uses a two-stage SFT+GRPO training pipeline with rewards for geo-accuracy, visual evidence grounding, and reasoning validity. The main empirical claim is that HoloGeo achieves 27.27% city-level accuracy on LandmarkBias-3K versus 16.83% for the Qwen2.5-VL-7B backbone and 23.57% for the strongest prior method, while maintaining competitive results on IM2GPS, IM2GPS3K, and YFCC4K.

Significance. If the results are correct, the paper would make a useful contribution: BI/BH provide a principled way to measure landmark over-reliance, LandmarkBias-3K targets a real failure mode, and the multi-reward GRPO framework is a plausible training strategy. The extensive qualitative analyses and the cross-model validation of BI/BH are additional strengths. However, the significance is conditional on resolving serious concerns about the construction of BF-30K and the definition of the BI/BH label space, both of which are load-bearing for the central claims.

major comments (4)
  1. [Appendix C, Fig. 9] The paper states in Appendix C that annotation models are 'without access to ground-truth location labels,' yet the BF-30K annotation prompt template in Figure 9 begins with 'The confirmed geographical location corresponding to the original image is: "country": {country}, "city": {city}' before asking the model to reason independently. This is a direct internal contradiction. Because these reasoning chains are the supervision signal for both SFT (§4.2) and the R_CLR reward (§4.3, Eq. 7), the claimed 'bias-free evidence-driven reasoning' is not established; the chains may be post-hoc rationalizations of the supplied label. The authors must resolve this contradiction and demonstrate that the training signal is not contaminated by the ground-truth label.
  2. [§3.1, Eqs. (1)–(4)] BI and BH are defined via probabilities P(y|x), P(y|x_r), and P(y|r), but the label space Y is never specified. It is unclear whether y ranges over all geographic entities, a fixed set of country/city names, or a closed vocabulary. Since the model is a generative VLM, the paper must define how P(y|x) is computed (e.g., token probabilities, constrained decoding, or ranking over a fixed list). Without this, the BI/BH values, the threshold selection in Appendix B (Eq. 10), and the resulting LandmarkBias-3K benchmark are not reproducible or verifiable.
  3. [§3.2 and §4.1] LandmarkBias-3K is constructed from MP-16 and GLDv2 (§3.2), and BF-30K is also filtered from MP-16 and GLDv2 (§4.1). The paper never states that the training set and the benchmark are disjoint. If there is overlap, the reported LandmarkBias-3K accuracy in Table 2 could be inflated by training-data memorization, which would undermine the central claim of systematic bias mitigation. The authors must confirm disjointness or quantify the overlap.
  4. [§4.3, Eq. (7)] The comprehensive logical reasoning reward R_CLR is computed by Qwen3.5-35B-A3B, a model from the same family as the policy backbone Qwen2.5-VL-7B. If the reward model shares the same landmark-bias tendencies as the policy, the RL loop may amplify rather than correct bias. The paper provides no analysis of the judge's reliability, agreement with human judgments, or sensitivity to the judge's own biases. At minimum, an ablation using a cross-family judge is needed to support the claim that R_CLR is 'comprehensive' and not circular.
minor comments (4)
  1. [Table 1] The GAEA row is formatted incorrectly: '43.0 57.477.236.9 56.073.2' should be separated as '43.0 57.4 77.2 | 36.9 56.0 73.2'. This makes the table hard to read.
  2. [Figure 1] The figure contains small text snippets (e.g., 'The Nessie statue suggests...') that are nearly illegible in the PDF. Please enlarge the font or remove the clutter.
  3. [Appendix F.3, Fig. 20] Several failure-case reasoning chains explicitly reference an 'original conclusion' (e.g., 'supports the original conclusion of Accra, Ghana'), which suggests the annotation prompts may have conditioned on prior answers. This is consistent with the label-leakage concern and should be clarified or corrected.
  4. [§3.1, Eq. (4)] The definition of BH combines positive and negative log-differences. A short worked example with concrete probability values would help readers interpret the sign and magnitude of BH.

Circularity Check

2 steps flagged · score 6.0 of 10

BF-30K annotation prompts supply ground-truth locations, contradicting the paper's 'no label access' claim; the evidence-driven training signal is post-hoc rationalization, and the bias benchmark is selected using the same Qwen BI/BH scoring used to filter the training data.

  1. self definitional [Appendix C (Detailed BF-30K Annotation) and Figure 9 prompt template]
    "To ensure faithful reasoning, the prompt (Figure 9) explicitly restricts the model to rely solely on visual evidence, without access to ground-truth location labels. ... [Figure 9 prompt:] The confirmed geographical location corresponding to the original image is: 'country': {country}, 'city': {city}. Your final task is to INDEPENDENTLY determine the country and city where the original image is located based solely on visual evidence."

    The annotation prompt gives the ground-truth country/city before requesting independent reasoning. The resulting <Analyze>/<Think> chains are therefore answer-conditioned rationalizations, not bias-free visual-evidence reasoning. Since BF-30K is the sole source of SFT supervision (Section 4.2) and the RCLR reward supervises reasoning consistency (Section 4.3), HoloGeo's claimed 'evidence-driven joint reasoning' reduces to imitating a model that already knew the answer. The paper's own Section C claim that labels were withheld is directly contradicted by Figure 9.

  2. self definitional [Sections 3.1-3.2 (LandmarkBias-3K construction) and Section 4.1 (BF-30K filtering)]
    "To ensure a stable probability estimation, Qwen2.5-VL-72B [2] is adopted for all probability estimates of I1, I2, and I3. ... For landmark-containing images, we further compute BI and BH scores to identify instances exhibiting strong landmark bias characteristics."

    The same Qwen2.5-VL-72B model and the same BI/BH scoring are used both to select LandmarkBias-3K (the bias benchmark) and to filter landmark-biased images into BF-30K (the training set). The evaluation therefore measures HoloGeo (a Qwen2.5-VL-7B model) on a benchmark that was chosen to be biased under the same scoring function that also curated its training data. Improvement on LandmarkBias-3K is partly a measure of fit to the Qwen-family bias signature, not an independent test of general landmark-bias mitigation.

full rationale

The strongest circularity is internal: Section C asserts the BF-30K annotation models reason 'without access to ground-truth location labels,' but Figure 9, the actual annotation prompt, begins with the confirmed country and city. This makes the 'bias-free, evidence-driven' reasoning chains post-hoc rationalizations, and because those chains supervise both SFT and RCLR, the central mechanism of HoloGeo is not established by the paper's own evidence. A second loop exists in the benchmark construction: both LandmarkBias-3K and the landmark-biased portion of BF-30K are selected using BI/BH computed with Qwen2.5-VL-72B, so the reported gain on LandmarkBias-3K is partially self-referential. The paper's standard-benchmark results (IM2GPS, IM2GPS3K, YFCC4K) are external and provide independent evidence that the training procedure is not wholly vacuous; hence the score is 6 rather than higher. No additional self-citation chain or uniqueness theorem is load-bearing.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The central claim depends on calibrated location-label probabilities, reliable landmark detection, and unbiased annotation. The first two are unstated domain assumptions; the third is contradicted by the appendix prompt, which is the main load-bearing flaw.

free parameters (6)
  • alpha (city-level weight in Rgeo) = 0.8
    Chosen to balance city vs country credit; not derived from data.
  • lambda1 (Rgeo reward weight) = 1.2
    Hand-set reward weight in Eq. 8.
  • lambda2 (Rbox reward weight) = 0.3
    Hand-set reward weight in Eq. 8.
  • lambda3 (RCLR reward weight) = 0.5
    Hand-set reward weight in Eq. 8.
  • tau_IoU (Rbox matching threshold) = 0.6
    IoU threshold for box coverage reward in Eq. 6.
  • BI/BH inclusion thresholds = BI > 0, BH > 0
    Selection criteria for LandmarkBias-3K; no principled threshold derivation is given.
assumptions (4)
  • domain assumption VLM output token probabilities can be interpreted as a calibrated distribution P(y|x) over all candidate location labels
    BI/BH and benchmark construction require this, but Sec 3.1 never specifies label space or probability extraction.
  • domain assumption GroundingDINO reliably detects the single prominent landmark region for every image
    Used to construct landmark-only and landmark-removed views; no failure analysis is provided.
  • domain assumption MP-16/GLDv2 labels are correct and sufficient for city-level geo-localization
    Used as ground truth for benchmark and training rewards.
  • ad hoc to paper BF-30K reasoning annotations were generated without seeing ground-truth labels
    Section C claims this, but Figure 9 prompt includes confirmed country/city; the paper depends on this to justify bias-free chains.

how reviews work

0 comments
Cite this review

Pith. "Pith review of HoloGeo: Mitigating Landmark Bias in Geo-localization via Evidence-Driven Reasoning." pith.science (2026). https://pith.science/paper/ATHHNWOH

@misc{pith2026260715255,
  author       = {Pith},
  title        = {Pith review of: HoloGeo: Mitigating Landmark Bias in Geo-localization via Evidence-Driven Reasoning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ATHHNWOH}},
  note         = {Machine review of arXiv:2607.15255}
}
read the original abstract

Recent advances in Vision-Language Models (VLMs) have significantly improved image geo-localization, yet existing models remain susceptible to landmark bias, causing them to overlook geographical cues or form spurious correlations, ultimately resulting in inaccurate localization. To systematically investigate this issue, we first design two quantitative metrics, Bias Intensity (BI) and Bias Harmfulness (BH), to characterize the impact of landmarks exerted on model reasoning, and establish a comprehensive benchmark, LandmarkBias-3K. To mitigate landmark bias, we further propose an evidence-driven reasoning framework, HoloGeo, to improve the reliability of geo-localization. HoloGeo is supported by a high-quality dataset, BF-30k, annotated with structured multi-evidence bias-free reasoning chains. By incorporating multi-dimensional rewards, HoloGeo explicitly encourages balanced attention over diverse visual cues and achieves evidence-driven joint reasoning. Extensive experiments demonstrate that HoloGeo not only maintains excellent performance on IM2GPS3K and YFCC4k but also significantly outperforms existing open-source VLMs on LandmarkBias-3K, validating its effectiveness for robust geospatial reasoning.

Figures

Figures reproduced from arXiv: 2607.15255 by the authors.

Figure 1
Figure 1. Illustration of landmark bias in geospatial reason [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Pipeline for constructing LandmarkBias-3K. For each image, landmark regions are first extracted to derive landmark [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Overview of the proposed HoloGeo. Stage I constructs the BF-30K dataset by filtering samples from MP-16 [ [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (17 more)
Figure 4
Figure 4. Figure 4: Comparison of Attention Distributions Across Qwen2.5-VL-7B-Instruct, GLOBE, GeoAgent, GeoReasoner, and HoloGeo. [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Prediction of Qwen2.5-VL-7B and HoloGeo on dif￾ferent landmark types. Romania, Bucharest HoloGeo Romania, Bucharest "Box": [628, 429, 1700, 825], "Analysis": "The building in the background has a classical architectural style, featuring columns, arches, and a prominent…
Figure 6
Figure 6. Figure 6: Case study: Reasoning comparison of three differ [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: Empirical distributions of Bias Intensity (BI) and Bias Harmfulness (BH) in LandmarkBias-3K. The dashed lines mark [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]
Figure 8
Figure 8. Figure 8: Representative cases with different BI/BH sign combinations: (a) [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]
Figure 9
Figure 9. Figure 9: The prompt template used in BF-30K annotation. [PITH_FULL_IMAGE:figures/full_fig_p013_9.png]
Figure 10
Figure 10. Figure 10: Representative cases of landmark categories in LandmarkBias-3K. [PITH_FULL_IMAGE:figures/full_fig_p014_10.png]
Figure 11
Figure 11. Figure 11: Visualization of an example of the training dataset. [PITH_FULL_IMAGE:figures/full_fig_p015_11.png]
Figure 12
Figure 12. Figure 12: The prompt template used in the InternVL3-78B [PITH_FULL_IMAGE:figures/full_fig_p015_12.png]
Figure 13
Figure 13. Figure 13: Prompts in SFT and GRPO. United Kingdom,Bristol <Analyze> {"Box": [368, 1557, 1549, 2340], "Analysis": "The architecture of the buildings in the foreground resembles Georgian-style residential housing commonly found in cities like Bath, England."}, {"Box": [1859, 64, …
Figure 14
Figure 14. Figure 14: A case of RCLR [PITH_FULL_IMAGE:figures/full_fig_p016_14.png]
Figure 15
Figure 15. Figure 15: Comparison of the BI and BH distribu￾tions between Qwen2.5-VL-7B-Instruct and HoloGeo on LandmarkBias-3K. based on coarse landmark-level similarity rather than robust geo￾graphic evidence. In contrast, our model remains less susceptible to such dominant visual bias: i…
Figure 16
Figure 16. Figure 16: Case study: Reasoning comparison of three different models on the same input image. Reliable geographic information [PITH_FULL_IMAGE:figures/full_fig_p018_16.png]
Figure 17
Figure 17. Figure 17: Cross-model validation of BH and BI based on tercile agreement between InternVL3-78B and Qwen2.5-VL-72B. The [PITH_FULL_IMAGE:figures/full_fig_p018_17.png]
Figure 18
Figure 18. Figure 18: More correct cases of HoloGeo on LandmarkBias-3K. [PITH_FULL_IMAGE:figures/full_fig_p019_18.png]
Figure 19
Figure 19. Figure 19: More correct cases of HoloGeo on IM2GPS. [PITH_FULL_IMAGE:figures/full_fig_p020_19.png]
Figure 20
Figure 20. Figure 20: Some failure cases of HoloGeo on IM2GPS. [PITH_FULL_IMAGE:figures/full_fig_p021_20.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

145 extracted references · 13 linked inside Pith

  1. [1]

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Floren- cia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. 2023. GPT-4 technical report.arXiv preprint arXiv:2303.08774 (2023)

    arXiv 2023

  2. [2]

    Shuai Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Sibo Song, Kai Dang, Peng Wang, Shijie Wang, Jun Tang, Humen Zhong, Yuanzhi Zhu, Mingkun Yang, Zhaohai Li, Jianqiang Wan, Pengfei Wang, Wei Ding, Zheren Fu, Yiheng Xu, Jiabo Ye, Xi Zhang, Tianbao Xie, Zesen Cheng, Hang Zhang, Zhibo Yang, Haiyang Xu, and Junyang Lin. 2025. Qwen2.5-VL Technical Rep...

    arXiv 2025

  3. [3]

    Maurits Bleeker, Mariya Hendriksen, Andrew Yates, and Maarten de Rijke. 2024. Demonstrating and Reducing Shortcuts in Vision-Language Representation Learning.Transactions on Machine Learning Research(2024)

    2024

  4. [4]

    Ron Campos, Ashmal Vayani, Parth Parag Kulkarni, Rohit Gupta, Aizan Zafar, Aritra Dutta, and Mubarak Shah. 2026. Gaea: A geolocation aware conversational assistant. InProceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision. 5236–5246

    2026

  5. [5]

    Athanasios Chalvatzaras, Ioannis Pratikakis, and Angelos A Amanatiadis. 2022. A survey on map-based localization techniques for autonomous vehicles.IEEE Transactions on intelligent vehicles8, 2 (2022), 1574–1596

    2022

  6. [6]

    Cong Chen, Mingyu Liu, Chenchen Jing, Yizhou Zhou, Fengyun Rao, Hao Chen, Bo Zhang, and Chunhua Shen. 2025. PerturboLLaVA: Reducing Multimodal Hallucinations with Perturbative Visual Training. InThe Thirteenth Interna- tional Conference on Learning Representations. https://openreview.net/forum?id= j4LITBSUjs

    2025

  7. [7]

    Zhe Chen, Jiannan Wu, Wenhai Wang, Weijie Su, Guo Chen, Sen Xing, Muyan Zhong, Qinglong Zhang, Xizhou Zhu, Lewei Lu, et al . 2024. Internvl: Scaling up vision foundation models and aligning for generic visual-linguistic tasks. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 24185–24198

    2024

  8. [8]

    Fenghua Cheng, Jinxiang Wang, Sen Wang, Zi Huang, and Xue Li. 2025. GeoGuess: Multimodal Reasoning based on Hierarchy of Visual Information in Street View. arXiv preprint arXiv:2506.16633(2025)

    arXiv 2025

Show all 145 references
  1. [9]

    Wenqing Cheng, Ruxue Wen, Haojun Huang, Wang Miao, and Chen Wang

  2. [10]

    Zihui Cheng, Qiguang Chen, Xiao Xu, Jiaqi Wang, Weiyun Wang, Hao Fei, Yidong Wang, Alex Jinpeng Wang, Zhi Chen, Wanxiang Che, et al. 2025. Visual thoughts: A unified perspective of understanding multimodal chain-of-thought.arXiv preprint arXiv:2505.15510(2025)

    2025

  3. [11]

    Ziheng Chi, Yifan Hou, Chenxi Pang, Shaobo Cui, Mubashara Akhtar, and Mrin- maya Sachan. 2025. Chimera: Diagnosing Shortcut Learning in Visual-Language Understanding.arXiv preprint arXiv:2509.22437(2025)

    2025

  4. [12]

    Tianzhe Chu, Yuexiang Zhai, Jihan Yang, Shengbang Tong, Saining Xie, Dale Schuurmans, Quoc V Le, Sergey Levine, and Yi Ma. 2025. SFT Memorizes, RL Generalizes: A Comparative Study of Foundation Model Post-training. InInter- national Conference on Machine Learning. PMLR, 10818–10838

    2025

  5. [13]

    Brandon Clark, Alec Kerrigan, Parth Parag Kulkarni, Vicente Vivanco Cepeda, and Mubarak Shah. 2023. Where we are and what we’re looking at: Query based worldwide image geo-localization using hierarchies and scenes. InProceedings of the IEEE/CVF Conference on Computer Vision an...

    2023

  6. [14]

    Zhiyang Dou, Zipeng Wang, Xumeng Han, Guorong Li, Zhipei Huang, and Zhenjun Han. 2024. Gaga: Towards interactive global geolocation assistant. arXiv preprint arXiv:2412.08907(2024)

    2024 arXiv

  7. [15]

    Nicolas Dufour, Vicky Kalogeiton, David Picard, and Loic Landrieu. 2025. Around the world in 80 timesteps: A generative approach to global visual geolocation. In Proceedings of the Computer Vision and Pattern Recognition Conference. 23016– 23026

    2025

  8. [16]

    Aritra Dutta, Srijan Das, Jacob Nielsen, Rajatsubhra Chakraborty, and Mubarak Shah. 2024. Multiview aerial visual recognition (mavrec): Can multi-view improve aerial visual perception?. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 22678–22690

    2024

  9. [17]

    Yu Feng, Puzhen Zhang, Guohui Xiao, Linfang Ding, and Liqiu Meng. 2025. Towards a barrier-free GeoQA portal: Natural language interaction with geospa- tial data using multi-agent LLMs and semantic search.Int. J. Appl. Earth Obs. Geoinformation144 (2025), 104825

    2025

  10. [18]

    Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, Ruoyu Zhang, Runxin Xu, Qihao Zhu, Shirong Ma, Peiyi Wang, Xiao Bi, et al . 2025. Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning.arXiv preprint arXiv:2501.12948(2025)

    2025 arXiv

  11. [19]

    Lukas Haas, Silas Alberti, and Michal Skreta. 2023. Learning generalized zero-shot learners for open-domain image geolocalization.arXiv preprint arXiv:2302.00275 (2023)

    2023 arXiv

  12. [20]

    Lukas Haas, Michal Skreta, Silas Alberti, and Chelsea Finn. 2024. Pigeon: Predict- ing image geolocations. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 12893–12902

    2024

  13. [21]

    Xiao Han, Chen Zhu, Hengshu Zhu, and Xiangyu Zhao. 2025. Swarm intelligence in geo-localization: A multi-agent large vision-language model collaborative framework. InProceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining V. 2. 814–825

    2025

  14. [22]

    James Hays and Alexei A Efros. 2008. Im2gps: estimating geographic informa- tion from a single image. In2008 ieee conference on computer vision and pattern recognition. 1–8

    2008

  15. [23]

    Edward J Hu, yelong shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. 2022. LoRA: Low-Rank Adaptation of Large Language Models. InInternational Conference on Learning Representations. https: //openreview.net/forum?id=nZeVKeeFYf9

    2022

  16. [24]

    Jie Huang, Xinyun Chen, Swaroop Mishra, Huaixiu Steven Zheng, Adams Wei Yu, Xinying Song, and Denny Zhou. 2024. Large Language Models Cannot Self-Correct Reasoning Yet. InThe Twelfth International Conference on Learning Representations

    2024

  17. [25]

    Pengyue Jia, Yiding Liu, Xiaopeng Li, Yuhao Wang, Yantong Du, Xiao Han, Xuetao Wei, Shuaiqiang Wang, Dawei Yin, and Xiangyu Zhao. 2024. G3: an effective and adaptive framework for worldwide geolocalization using large multi-modality models.Advances in Neural Information Proces...

    2024

  18. [26]

    Pengyue Jia, Seongheon Park, Song Gao, Xiangyu Zhao, and Yixuan Li. 2025. GeoRanker: Distance-Aware Ranking for Worldwide Image Geolocalization.arXiv preprint arXiv:2505.13731(2025)

    2025 arXiv

  19. [27]

    Modi Jin, Yiming Zhang, Boyuan Sun, Dingwen Zhang, Ming-Ming Cheng, and Qibin Hou. 2026. Geoagent: Learning to geolocate everywhere with reinforced geographic characteristics. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 41352–41364

    2026

  20. [28]

    Kartik Kuckreja, Muhammad Sohail Danish, Muzammal Naseer, Abhijit Das, Salman Khan, and Fahad Shahbaz Khan. 2024. Geochat: Grounded large vision- language model for remote sensing. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition. 27831–27840

    2024

  21. [29]

    Martha Larson, Mohammad Soleymani, Guillaume Gravier, Bogdan Ionescu, and Gareth JF Jones. 2017. The benchmarking initiative for multimedia evaluation: MediaEval 2016.IEEE MultiMedia24, 1 (2017), 93–96

    2017

  22. [30]

    Bin Li, Dehong Gao, Yeyuan Wang, Linbo Jin, Shanqing Yu, Xiaoyan Cai, and Libin Yang. 2025. Instruction-Aligned Visual Attention for Mitigating Hallucina- tions in Large Vision-Language Models. In2025 IEEE International Conference on Multimedia and Expo (ICME). IEEE, 1–6

    2025

  23. [31]

    Ling Li, Yu Ye, Bingchuan Jiang, and Wei Zeng. 2024. GeoReasoner: Geo- localization with Reasoning in Street Views using a Large Vision-Language Model. InInternational Conference on Machine Learning. PMLR, 29222–29233

    2024

  24. [32]

    Lingyao Li, Runlong Yu, Qikai Hu, Bowei Li, Min Deng, Yang Zhou, and Xiaowei Jia. 2025. From pixels to places: A systematic benchmark for evaluating image geolocalization ability in large language models.arXiv preprint arXiv:2508.01608 (2025). 9 MM ’26, November 10–14, 2026, R...

    2025 arXiv

  25. [33]

    Ling Li, Yao Zhou, Yuxuan Liang, Fugee Tsung, and Jiaheng Wei. 2025. Recogni- tion through Reasoning: Reinforcing Image Geo-localization with Large Vision- Language Models. InAdvances in Neural Information Processing Systems

    2025

  26. [34]

    Wen Li, Limin Wang, Wei Li, Eirikur Agustsson, and Luc Van Gool. 2017. Webvi- sion database: Visual learning and understanding from web data.arXiv preprint arXiv:1708.02862(2017)

    2017 arXiv

  27. [35]

    Yuxuan Liang, Songyu Ke, Junbo Zhang, Xiuwen Yi, and Yu Zheng. 2018. Geoman: Multi-level attention networks for geo-sensory time series prediction.. InIjcai, Vol. 2018. 3428–3434

    2018

  28. [36]

    Haotian Liu, Chunyuan Li, Yuheng Li, Bo Li, Yuanhan Zhang, Sheng Shen, and Yong Jae Lee. 2024. LLaVA-NeXT: Improved reasoning, OCR, and world knowl- edge

    2024

  29. [37]

    Jiazhen Liu, Yuhan Fu, Ruobing Xie, Runquan Xie, Xingwu Sun, Fengzong Lian, Zhanhui Kang, and Xirong Li. 2025. PhD: A ChatGPT-Prompted Visual Hallucina- tion Evaluation Dataset. InIEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2025, Nashville, TN, USA, Ju...

    2025

  30. [38]

    Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Qing Jiang, Chunyuan Li, Jianwei Yang, Hang Su, et al. 2024. Grounding dino: Marry- ing dino with grounded pre-training for open-set object detection. InEuropean conference on computer vision

    2024

  31. [39]

    Yi Liu, Junchen Ding, Gelei Deng, Yuekang Li, Tianwei Zhang, Weisong Sun, Yaowen Zheng, Jingquan Ge, and Yang Liu. 2024. Image-based geolocation using large vision-language models.arXiv preprint arXiv:2408.09474(2024)

    2024 arXiv

  32. [40]

    Yunze Man, De-An Huang, Guilin Liu, Shiwei Sheng, Shilong Liu, Liang-Yan Gui, Jan Kautz, Yu-Xiong Wang, and Zhiding Yu. 2025. Argus: Vision-Centric Reasoning with Grounded Chain-of-Thought. InProceedings of the Computer Vision and Pattern Recognition Conference. 14268–14280

    2025

  33. [41]

    Tessa Masis and Brendan O’Connor. 2026. Coordinates from Context: Using LLMs to Ground Complex Location References. InProceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers). 1591–1606

    2026

  34. [42]

    Eric Muller-Budack, Kader Pustu-Iren, and Ralph Ewerth. 2018. Geolocation estimation of photos using a hierarchical model and scene classification. In Proceedings of the European conference on computer vision (ECCV). 563–579

    2018

  35. [43]

    OpenAI. 2024. GPT-4o mini: Advancing Cost-Efficient Intelligence. https:// openai.com/index/gpt-4o-mini-advancing-cost-efficient-intelligence/

    2024

  36. [44]

    OpenAI. 2024. GPT-4o System Card. https://openai.com/index/gpt-4o-system- card/

    2024

  37. [45]

    OpenAI. 2025. Introducing o3 and o4-mini. https://openai.com/index/introducing- o3-and-o4-mini/

    2025

  38. [46]

    Zheyi Pan, Yuxuan Liang, Weifeng Wang, Yong Yu, Yu Zheng, and Junbo Zhang

  39. [47]

    Nhi Pham and Michael Schott. 2024. H-POPE: Hierarchical Polling-based Probing Evaluation of Hallucinations in Large Vision-Language Models. InWorkshop on Statistical Foundations of LLMs and Foundation Models

    2024

  40. [48]

    Shraman Pramanick, Ewa M Nowara, Joshua Gleason, Carlos D Castillo, and Rama Chellappa. 2022. Where in the world is this image? transformer-based geo-localization in the wild. InEuropean Conference on Computer Vision. Springer, 196–215

    2022

  41. [49]

    Qwen Team. 2026. Qwen3.5: Towards Native Multimodal Agents. https://qwen. ai/blog?id=qwen3.5

    2026

  42. [50]

    Paul Hongsuck Seo, Tobias Weyand, Jack Sim, and Bohyung Han. 2018. CPlaNet: Enhancing Image Geolocalization by Combinatorial Partitioning of Maps. In Proceedings of the European Conference on Computer Vision

    2018

  43. [51]

    Qiaomu Shen, Wei Zeng, Yu Ye, Stefan Müller Arisona, Simon Schubiger, Remo Burkhard, and Huamin Qu. 2017. StreetVizor: Visual exploration of human- scale urban forms based on street views.IEEE transactions on visualization and computer graphics24, 1 (2017), 1004–1013

    2017

  44. [52]

    Bart Thomee, David A Shamma, Gerald Friedland, Benjamin Elizalde, Karl Ni, Douglas Poland, Damian Borth, and Li-Jia Li. 2016. Yfcc100m: The new data in multimedia research.Commun. ACM(2016)

    2016

  45. [53]

    Vicente Vivanco Cepeda, Gaurav Kumar Nayak, and Mubarak Shah. 2023. Geoclip: Clip-inspired alignment between locations and images for effective worldwide geo-localization.Advances in Neural Information Processing Systems36 (2023), 8690–8701

    2023

  46. [54]

    An Vo, Khai-Nguyen Nguyen, Mohammad Reza Taesiri, Vy Tuong Dang, Anh Totti Nguyen, and Daeyoung Kim. 2025. Vision language models are biased. arXiv preprint arXiv:2505.23941(2025)

    2025 arXiv

  47. [55]

    Nam Vo, Nathan Jacobs, and James Hays. 2017. Revisiting im2gps in the deep learning era. InProceedings of the IEEE international conference on computer vision. 2621–2630

    2017

  48. [56]

    Beining Wang, Weihang Su, Hongtao Tian, Tao Yang, Yujia Zhou, Ting Yao, Qingyao Ai, and Yiqun Liu. 2025. From< Answer> to< Think>: Multidimen- sional Supervision of Reasoning Process for LLM Optimization.arXiv preprint arXiv:2510.11457(2025)

    2025

  49. [57]

    Chun Wang, Xiaojun Ye, Xiaoran Pan, Zihao Pan, Haofan Wang, and Yiren Song. 2025. Gre suite: Geo-localization inference via fine-tuned vision-language models and enhanced reasoning chains.Advances in Neural Information Processing Systems(2025)

    2025

  50. [58]

    Xiaolong Wang, Runsen Xu, Zhuofan Cui, Zeyu Wan, and Yu Zhang. 2023. Fine- grained cross-view geo-localization using a correlation-aware homography esti- mator.Advances in Neural Information Processing Systems36 (2023), 5301–5319

    2023

  51. [59]

    Jiaheng Wei, Zhaowei Zhu, Hao Cheng, Tongliang Liu, Gang Niu, and Yang Liu

  52. [60]

    Jiaheng Wei, Zhaowei Zhu, Tianyi Luo, Ehsan Amid, Abhishek Kumar, and Yang Liu. 2023. To aggregate or not? learning with separate noisy labels. InProceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. 2523–2535

    2023

  53. [61]

    Weyand, A

    T. Weyand, A. Araujo, B. Cao, and J. Sim. 2020. Google Landmarks Dataset v2 - A Large-Scale Benchmark for Instance-Level Recognition and Retrieval. InProc. CVPR

    2020

  54. [62]

    InInternational Conference on Learning Representations

    Learning with Noisy Labels Revisited: A Study Using Real-World Human Annotations. InInternational Conference on Learning Representations

  55. [63]

    Wikimedia Foundation. [n. d.]. Wikimedia API. https://commons.wikimedia.org/ w/api.php

  56. [64]

    Jiayang Wu, Wensheng Gan, Han-Chieh Chao, and Philip S. Yu. 2024. Geospatial Big Data: Survey and Challenges.IEEE J. Sel. Top. Appl. Earth Obs. Remote. Sens. 17 (2024), 17007–17020

    2024

  57. [65]

    Tobias Weyand, Ilya Kostrikov, and James Philbin. 2016. Planet-photo geolocation with convolutional neural networks. InEuropean conference on computer vision

    2016

  58. [66]

    Zimin Xia and Alexandre Alahi. 2025. FGˆ 2: Fine-Grained Cross-View Localiza- tion by Fine-Grained Feature Matching. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 6362–6372

    2025

  59. [67]

    Chenhui Xu, Fuxun Yu, Michael J. Bianco, Jacob Kovarskiy, Raphael Tang, Qi Zhang, Zirui Xu, Will LeVine, Brandon Dubbs, Heming Liao, Cassandra Burgess, Suvam Bag, Jay Patravali, Rupanjali Kukal, Mikael Figueroa, Rishi Madhok, Niko- laos Karianakis, and Jinjun Xiong. 2026. Unlo...

    2026

  60. [68]

    Shengqiong Wu, Hao Fei, Liangming Pan, William Yang Wang, Shuicheng Yan, and Tat-Seng Chua. 2025. Combating multimodal llm hallucination via bottom-up holistic reasoning. InProceedings of the AAAI Conference on Artificial Intelligence, Vol. 39. 8460–8468

    2025

  61. [69]

    Hongji Yang, Xiufan Lu, and Yingying Zhu. 2021. Cross-view geo-localization with layer-to-layer transformer.Advances in Neural Information Processing Sys- tems34 (2021), 29009–29020

    2021

  62. [70]

    Sahiti Yerramilli, Nilay Pande, Rynaa Grover, and Jayant Sravan Tamarapalli

  63. [71]

    Yibo Yan, Haomin Wen, Siru Zhong, Wei Chen, Haodong Chen, Qingsong Wen, Roger Zimmermann, and Yuxuan Liang. 2024. UrbanCLIP: Learning Text- enhanced Urban Region Profiling with Contrastive Language-Image Pretraining from the Web. InProceedings of the ACM on Web Conference 2024...

    2024

  64. [72]

    Zilun Zhang, Zian Guan, Tiancheng Zhao, Haozhan Shen, Tianyu Li, Yuxiang Cai, Zhonggen Su, Zhaojun Liu, Jianwei Yin, and Xiang Li. 2025. Geo-R1: Improving Few-Shot Geospatial Referring Expression Understanding with Reinforcement Fine-Tuning.arXiv preprint arXiv:2509.21976(2025)

    2025 arXiv

  65. [73]

    Lianmin Zheng, Wei-Lin Chiang, Ying Sheng, Siyuan Zhuang, Zhanghao Wu, Yonghao Zhuang, Zi Lin, Zhuohan Li, Dacheng Li, Eric Xing, et al. 2023. Judging llm-as-a-judge with mt-bench and chatbot arena.Advances in neural information processing systems36 (2023), 46595–46623

    2023

  66. [74]

    Chunting Zhou, Pengfei Liu, Puxin Xu, Srinivasan Iyer, Jiao Sun, Yuning Mao, Xuezhe Ma, Avia Efrat, Ping Yu, Lili Yu, et al . 2023. Lima: Less is more for alignment.Advances in Neural Information Processing Systems36 (2023), 55006– 55021

    2023

  67. [75]

    Xintong Zhang, Zhi Gao, Bofei Zhang, Pengxiang Li, Xiaowen Zhang, Yang Liu, Tao Yuan, Yuwei Wu, Yunde Jia, Song-Chun Zhu, et al. 2025. Chain-of-Focus: Adaptive Visual Search and Zooming for Multimodal Reasoning via RL.arXiv preprint arXiv:2505.15436(2025)

    2025 arXiv

  68. [76]

    Jinguo Zhu, Weiyun Wang, Zhe Chen, Zhaoyang Liu, Shenglong Ye, Lixin Gu, Hao Tian, Yuchen Duan, Weijie Su, Jie Shao, et al. 2025. Internvl3: Exploring advanced training and test-time recipes for open-source multimodal models. arXiv preprint arXiv:2504.10479(2025)

    2025 arXiv

  69. [77]

    Box": [x0, y0, x1, y1],

    Sijie Zhu, Mubarak Shah, and Chen Chen. 2022. Transgeo: Transformer is all you need for cross-view image geo-localization. InProceedings of the IEEE/CVF 10 HoloGeo: Mitigating Landmark Bias in Geo-localization via Evidence-Driven Reasoning MM ’26, November 10–14, 2026, Rio de ...

    2022

  70. [79]

    Zhongliang Zhou, Jielu Zhang, Zihan Guan, Mengxuan Hu, Ni Lao, Lan Mu, Sheng Li, and Gengchen Mai. 2024. Img2Loc: Revisiting image geolocalization using multi-modality foundation models and image-based retrieval-augmented generation. InProceedings of the 47th international acm...

    2024

  71. [82]

    In the <Analyze> section: All describedvisual features are consistent with the actualcontent in the image

  72. [83]

    In the <Think> section: The reasoning logicis coherent, and does not rely solely on a singlegeographic feature unless no other identifiableobjects or visual clues exist in the image

  73. [84]

    Box": [368, 1557, 1549, 2340],

    In the <Answer> section: The country andcity fields are filled with specific, concrete placenames, and do NOT contain 'Not specified','Unknown', or other vague expressions. Criteria for 'no': If ANY of the above three conditions is not met,output 'no'. Note: Only output 'yes' ...

  74. [85]

    UK" to "Bath

    In the <Analyze> section: The analysis correctly identifies key visualelements present in the image: the Georgian-style building, the GothicRevival church tower, the red telephone box, and the bare trees indicatingwinter. These descriptions are factually consistent with the pr...

    2026

  75. [86]

    Core Evidence Synthesis: Medieval fortifications, Orthodox religious architecture with green domes, an Eastern European settlement layout, and a rural landscape typical of Ukraine together indicate strong Ukrainian cultural heritage

  76. [87]

    Among Ukrainian landmarks, these features most specifically match Khotyn Fortress, which is well known for its historical significance and preserved cultural setting

    Logical Reasoning Chain: The combination of medieval fortress design, Orthodox religious architecture, and a rural Eastern European settlement pattern points to Ukraine. Among Ukrainian landmarks, these features most specifically match Khotyn Fortress, which is well known for ...

  77. [88]

    Conflict Resolution: There are no conflicting visual information points, as all observed evidence consistently aligns with the geographical and cultural context of Ukraine

  78. [89]

    Box": [68, 107, 169, 520],

    Conclusion Justification: The full set of architectural, cultural, and landscape cues strongly supports identifying the location as Khotyn Fortress in Ukraine. The image shows a medieval fortress with three pointed towers, situated on a hillside overlooking a river. The fortre...

    2026

  79. [90]

    Urban Context: The coexistence of historical religious structures and modern buildings indicates a contemporary urban environment

    Core Evidence Synthesis: Architectural Style: The churches display Russian Orthodox architectural elements, including red and white colors, onion-shaped domes, and crosses. Urban Context: The coexistence of historical religious structures and modern buildings indicates a conte...

  80. [91]

    The surrounding modern residential and commercial buildings indicate a large contemporary city

    Logical Reasoning Chain: The church architecture points to a Russian cultural setting. The surrounding modern residential and commercial buildings indicate a large contemporary city. The combination of historic religious landmarks and modern urban development is most consisten...

  81. [92]

    Conflict Resolution: There are no conflicting visual cues across the boxes

  82. [93]

    Box": [0, 0, 367, 318],

    Conclusion Justification: The consistent Russian Orthodox architectural style and the mixed historic-modern urban setting strongly suggest Moscow, Russia. HoloGeo Italy, Venice "Box": [0, 0, 367, 318], "Analysis": "The image shows a canal with historic buildings featuring bric...

  83. [94]

    Canal Scene: The canal, boats, and railings indicate a water-adapted setting

    Core Evidence Synthesis: Architecture: Brick facades, small windows, and decorative details match Venetian design. Canal Scene: The canal, boats, and railings indicate a water-adapted setting. Historical Context: Weathered buildings reflect historical heritage. Geographic Feat...

  84. [95]

    The canal and boats indicate a city centered on waterways

    Logical Reasoning Chain: The architecture suggests a historic Italian setting. The canal and boats indicate a city centered on waterways. Together, these cues are most consistent with Venice, Italy

  85. [96]

    Conflict Resolution: No conflicting visual cues are present

  86. [97]

    Box": [67, 139, 138, 317],

    Conclusion Justification: The canal setting and Venetian-style historic architecture strongly suggest Venice, Italy. HoloGeo HoloGeo Japan, Kyoto "Box": [67, 139, 138, 317], "Analysis": "The red lettering resembling Japanese characters suggests a Japanese commercial or cultura...

  87. [98]

    Architecture: The surrounding buildings reflect urban Japanese design

    Core Evidence Synthesis: Landmark: Kyoto Tower is a strong indicator of Kyoto. Architecture: The surrounding buildings reflect urban Japanese design. Infrastructure: The bus stop and signage are consistent with Japanese urban infrastructure. Geographic Feature Correlation: The...

  88. [99]

    The architectural style and urban infrastructure support a Japanese city setting

    Logical Reasoning Chain: Kyoto Tower suggests Kyoto directly. The architectural style and urban infrastructure support a Japanese city setting. Together, these cues are most consistent with Kyoto, Japan

  89. [100]

    Box": [798, 326, 1420, 2008],

    Conclusion Justification: The presence of Kyoto Tower and the surrounding Japanese urban environment strongly suggest Kyoto, Japan. Ukraine, Kremenchuk HoloGeo "Box": [798, 326, 1420, 2008], "Analysis": "The left tower displays the text 'СЛАВСЯ УКРАЇНО' together with the Ukrai...

    2008

  90. [101]

    Place Name: The banner directly mentions Kremenchuk, providing a specific city cue

    Core Evidence Synthesis: Text and Symbols: The Ukrainian-language slogan and the national trident symbol strongly indicate Ukraine. Place Name: The banner directly mentions Kremenchuk, providing a specific city cue. Scene Context: The bridge entrance and passing vehicles sugge...

  91. [102]

    The explicit reference to Kremenchuk identifies the city directly

    Logical Reasoning Chain: The Ukrainian text and trident symbol establish the national context. The explicit reference to Kremenchuk identifies the city directly. The bridge setting and traffic further support that this is a real urban location rather than a symbolic display. T...

  92. [103]

    Box": [68, 179, 1806, 2368],

    Conclusion Justification: The combination of Ukrainian national symbols, Ukrainian-language text, and the direct mention of Kremenchuk strongly supports the conclusion that the image is from Kremenchuk, Ukraine. France, Les Mathes "Box": [68, 179, 1806, 2368], "Analysis": "The...

  93. [104]

    Place Name: The sign Les Mathes La Palmyre Zoo' directly points to Les Mathes

    Core Evidence Synthesis: Architecture: The church’s Gothic style is consistent with French religious architecture. Place Name: The sign Les Mathes La Palmyre Zoo' directly points to Les Mathes. Scene Context: The surrounding buildings and vehicles suggest a maintained rural co...

  94. [105]

    The sign provides a direct location clue

    Logical Reasoning Chain: The Gothic church suggests a historic French setting. The sign provides a direct location clue. The rural surroundings further support Les Mathes, France

  95. [106]

    Box": [498, 0, 1867, 1288],

    Conclusion Justification: The French-style church and the direct mention of Les Mathes strongly suggest Les Mathes, France. HoloGeo Spain, Girona "Box": [498, 0, 1867, 1288], "Analysis": "The pink brickwork and decorative stone facade suggest a Mediterranean architectural styl...

  96. [107]

    Environment: The ivy-covered wall and outdoor dining area add to the impression of a preserved, tourist-friendly historic town

    Core Evidence Synthesis: Architecture: The warm-toned brickwork, decorative stone facade, arches, and stone walls suggest a historic Mediterranean urban setting. Environment: The ivy-covered wall and outdoor dining area add to the impression of a preserved, tourist-friendly hi...

  97. [108]

    The stone construction and arched entryways support a medieval European context

    Logical Reasoning Chain: The building materials and facade details point to a Mediterranean historic city. The stone construction and arched entryways support a medieval European context. The outdoor seating and overall atmosphere fit a well-preserved Spanish old town. Togethe...

  98. [109]

    Box": [0, 0, 1036, 756],

    Conclusion Justification: The combination of Mediterranean-style architecture, medieval stone elements, and the historic urban atmosphere strongly suggests Girona, Spain. HoloGeo Figure 18: More correct cases of HoloGeo on LandmarkBias-3K. 19 MM ’26, November 10–14, 2026, Rio ...

    2026

  99. [110]

    Core Evidence Synthesis: English signage, urban building style, heavy traffic, locally common vehicles, and tropical vegetation together indicate a dense city environment in India

  100. [111]

    The traffic, vehicles, and architecture point to a large Indian city

    Logical Reasoning Chain: The English signage suggests a prominently English-using urban setting. The traffic, vehicles, and architecture point to a large Indian city. The greenery and tropical trees suggest southern India. Together, these cues are consistent with Bangalore

  101. [112]

    Conflict Resolution: There are no major conflicting visual cues; the evidence consistently supports an urban Indian location with a warm climate

  102. [113]

    country":

    Conclusion Justification: The combination of signage, traffic patterns, vehicle types, architecture, and vegetation supports the conclusion that the image is from India, with Bangalore as the proposed city. "country": "India", "city": "Bangalore" HoloGeo Bosnia and Herzegovina...

  103. [114]

    Core Evidence Synthesis: The main cues are medieval-style fortifications, traditional stone houses with Balkan features, and a visible minaret indicating Islamic cultural influence

  104. [115]

    The minaret and Ottoman-style architecture further narrow it to Bosnia and Herzegovina

    Logical Reasoning Chain: The fortress and stone residential buildings point to a historic Balkan town. The minaret and Ottoman-style architecture further narrow it to Bosnia and Herzegovina. Together, these features are consistent with Mostar

  105. [116]

    Conflict Resolution: There are no major conflicts among the visual clues; the architectural elements support the same regional interpretation

  106. [117]

    country":

    Conclusion Justification: The combination of historic fortification, Ottoman-influenced buildings, and Islamic religious architecture supports the conclusion that the image is from Bosnia and Herzegovina, with Mostar as the proposed city. "country": "Bosnia and Herzegovina", "...

  107. [118]

    Core Evidence Synthesis: The main cues are traditional European facades, German-language signage, the 'Hotel Goldener Hirsch' sign, and nearby mountains

  108. [119]

    The mountains narrow it toward Austria

    Logical Reasoning Chain: The architecture and signage point to a historic German-speaking European city. The mountains narrow it toward Austria. The 'Hotel Goldener Hirsch' sign makes Salzburg a plausible match

  109. [120]

    Conflict Resolution: There are no major conflicting visual cues; the architectural, linguistic, and geographic elements support the same interpretation

  110. [121]

    country":

    Conclusion Justification: The combination of historic European architecture, German signage, and mountainous surroundings supports the conclusion that the image is from Austria, with Salzburg as the proposed city. "country": "Austria", "city": "Salzburg" HoloGeo United Kingdom...

  111. [122]

    Core Evidence Synthesis: The main cues are the modern urban skyline, busy street traffic, and the presence of iconic taxis associated with London

  112. [123]

    The distinctive taxis strongly narrow it to London

    Logical Reasoning Chain: The mix of modern architecture and dense traffic points to a major global city. The distinctive taxis strongly narrow it to London. Together, these features are consistent with central London

  113. [124]

    Conflict Resolution: There are no major conflicting visual cues; the transportation and cityscape elements support the same interpretation

  114. [125]

    country":

    Conclusion Justification: The combination of recognizable taxis, urban traffic, and London-style architecture supports the conclusion that the image is from the United Kingdom, with London as the proposed city. "country": "United Kingdom", "city": "London" Portugal, Porto Holo...

  115. [126]

    Core Evidence Synthesis: The main cues are the traditional rabelo boats, the distinctive steel arch of the Dom Luís I Bridge, and the historic cityscape along the river

  116. [127]

    The bridge is a strong landmark cue for Porto

    Logical Reasoning Chain: The boats point to the Douro River region and Porto’s wine-transport heritage. The bridge is a strong landmark cue for Porto. The surrounding historic urban landscape further matches the city

  117. [128]

    Conflict Resolution: There are no major conflicting visual cues; the boats, bridge, and cityscape all support the same location

  118. [129]

    country":

    Conclusion Justification: The combination of rabelo boats, the Dom Luís I Bridge, and the riverside historic setting supports the conclusion that the image is from Portugal, with Porto as the proposed city. "country": "Portugal", "city": "Porto" Russia, Saint Petersburg HoloGe...

  119. [130]

    Core Evidence Synthesis: The main cues are the neoclassical architecture, cohesive historic urban layout, presence of buses, and everyday street activity

  120. [131]

    The building style is consistent with Russian urban centers, and the overall scene supports a large Russian city

    Logical Reasoning Chain: The architecture and city layout point to a historic European-style city. The building style is consistent with Russian urban centers, and the overall scene supports a large Russian city. Together, these features are consistent with Saint Petersburg

  121. [132]

    Conflict Resolution: There are no major conflicting visual cues; the architecture, transport, and street scene support the same interpretation

  122. [133]

    country":

    Conclusion Justification: The combination of neoclassical buildings, urban layout, and street activity supports the conclusion that the image is from Russia, with Saint Petersburg as the proposed city. "country": "Russia", "city": "Saint Petersburg" Figure 19: More correct cas...

    2026

  123. [134]

    Produce: Bananas, papayas, plantains, and vegetables indicate a tropical climate

    Core Evidence Synthesis: Market Scene: The open-air market, wooden stalls, and active commerce suggest a local community market. Produce: Bananas, papayas, plantains, and vegetables indicate a tropical climate. Environment: The tarpaulin coverings and informal setup suggest a ...

  124. [135]

    The informal market structure and community-based commerce are consistent with smaller towns or cities in developing regions

    Logical Reasoning Chain: The tropical produce suggests a warm climate. The informal market structure and community-based commerce are consistent with smaller towns or cities in developing regions. Together, these cues point broadly to sub-Saharan Africa, with the original reas...

  125. [136]

    Box": [0, 221, 416, 997],

    Conclusion Justification: The tropical produce, informal market layout, and overall atmosphere support the original conclusion of Accra, Ghana, though the evidence is more region-level than city-specific. HoloGeo Antigua and Barbuda, English Harbour HoloGeo "Box": [0, 221, 416...

  126. [137]

    Landscape: The rocky ground, sparse vegetation, and cliff-like coastal edge suggest a dry natural environment rather than a dense urban or forested area

    Core Evidence Synthesis: Vegetation: The agave plant is a strong indicator of an arid or semi-arid climate. Landscape: The rocky ground, sparse vegetation, and cliff-like coastal edge suggest a dry natural environment rather than a dense urban or forested area. Ocean Scene: Th...

  127. [138]

    The rocky terrain and coastal cliffs suggest a volcanic or island-like shoreline rather than an inland setting

    Logical Reasoning Chain: The agave and sparse vegetation point to a dry, sunny climate. The rocky terrain and coastal cliffs suggest a volcanic or island-like shoreline rather than an inland setting. The combination of arid vegetation, ocean views, and rugged coast is consiste...

  128. [139]

    Box": [0, 867, 755, 924],

    Conclusion Justification: The agave, rocky coastal terrain, sparse vegetation, and ocean scenery support the original conclusion of Santa Cruz de Tenerife, Spain, although the evidence is stronger at the regional or island-environment level than at the exact city level. Domini...

  129. [140]

    Environment: The bright sky, tropical vegetation, and lush misty mountains indicate a humid tropical climate

    Core Evidence Synthesis: Settlement: The modest houses, open fields, and simple residential structures suggest a rural or semi-rural community. Environment: The bright sky, tropical vegetation, and lush misty mountains indicate a humid tropical climate. Landscape: The mountain...

  130. [141]

    The lush green mountains and misty peaks point to a humid island landscape rather than a flat mainland environment

    Logical Reasoning Chain: The rural housing and warm-climate architecture suggest a small settlement in a tropical region. The lush green mountains and misty peaks point to a humid island landscape rather than a flat mainland environment. Together, these cues support the origin...

  131. [142]

    Box": [448, 67, 819, 596],

    Conclusion Justification: The rural settlement, tropical vegetation, and misty mountainous backdrop support the original conclusion of Roseau, Dominica, though the evidence is stronger at the island or regional level than at the exact city level. HoloGeo Montenegro, Herceg Nov...

  132. [143]

    Environment: The outdoor seating, umbrellas, greenery, and potted plants indicate a tourist-friendly coastal atmosphere

    Core Evidence Synthesis: Architecture: The arched windows, balconies, stone construction, and pinkish facade suggest a Mediterranean historic setting with Venetian influence. Environment: The outdoor seating, umbrellas, greenery, and potted plants indicate a tourist-friendly c...

  133. [144]

    The outdoor dining setup and decorative greenery support a popular tourist destination

    Logical Reasoning Chain: The stone buildings and Mediterranean design suggest a historic coastal town. The outdoor dining setup and decorative greenery support a popular tourist destination. Together, these cues support the original conclusion of Hvar, Croatia

  134. [145]

    Figure 20: Some failure cases of HoloGeo on IM2GPS

    Conclusion Justification: The Mediterranean architecture, Venetian-influenced details, and tourist-oriented coastal atmosphere support the original conclusion of Hvar, Croatia. Figure 20: Some failure cases of HoloGeo on IM2GPS. 21

  135. [2019]

    InProceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, KDD 2019, Anchorage, AK, USA, August 4-8,

    Urban Traffic Prediction from Spatio-Temporal Data Using Deep Meta Learning. InProceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, KDD 2019, Anchorage, AK, USA, August 4-8,

    2019

  136. [2022]

    OPTDP: Towards optimal personalized trajectory differential privacy for trajectory data publishing.Neurocomputing472 (2022), 201–211

    2022

  137. [2025]

    In Findings of the Association for Computational Linguistics: EMNLP 2025, Christos Christodoulopoulos, Tanmoy Chakraborty, Carolyn Rose, and Violet Peng (Eds.)

    GeoChain: Multimodal Chain-of-Thought for Geographic Reasoning. In Findings of the Association for Computational Linguistics: EMNLP 2025, Christos Christodoulopoulos, Tanmoy Chakraborty, Carolyn Rose, and Violet Peng (Eds.). Association for Computational Linguistics, Suzhou, C...

    2025

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.