Pith. sign in

REVIEW 3 major objections 7 minor 1 cited by

Multi-Agent Geospatial Copilots for Remote Sensing Workflows

T0 review · 3 major / 7 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read GeoLLM-Squad, a multi-agent geospatial copilot that separates planning from task-solving, achieves 60.29% correct tool-calling on remote sensing workflows, 17 points above a single-agent baseline.

desk verdict A sensible hybrid multi-agent architecture for RS workflows, but the 17% correctness claim is only as reliable as the GPT-4o-generated gold labels it is scored against. read the letter →

arxiv 2501.16254 v1 pith:ZE57MP73 submitted 2025-01-27 cs.LG

Chaehong Lee , Varatheepan Paramanayakam , Andreas Karatzas , Yanan Jian , Michael Fore , Heming Liao , Fuxun Yu , Ruopu Li
show 2 more authors
Iraklis Anagnostopoulos Dimitrios Stamoulis
This is my paper · ORCID
classification cs.LG
keywords geospatialcopilotsmulti-agentsystemsremotesensingworkflowsLLMagentstoolselectionworkflowmemoryagenticcorrectnesssatellitevision
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 argues that the reason geospatial AI assistants plateau is that a single large language model tries to do everything—load data, reason, call tools, and plot—inside one context window. It proposes GeoLLM-Squad, which separates orchestration from task-solving by delegating to specialized sub-agents, and adds tool-selection hints plus workflow memory. On a 2,000-task benchmark across agriculture, climate, urban, forest, and satellite-vision workflows, the system reaches 60.29% correct tool-calling, about 17 points above the single-agent baseline, at 78.49k tokens per task. If this holds, multi-agent architecture rather than a bigger monolithic model is the path to scaling remote sensing copilots.

What carries the argument

The load-bearing mechanism is a hybrid orchestrator: a supervisor agent decomposes the user request into subtasks, writes natural-language schedules that fix agent order (for example, load before filter), dispatches to specialized agents with dedicated toolkits that total 521 API functions, aggregates return messages, and if the task is incomplete revises the schedule and loops. Two augmentation techniques carry the few-shot signal: intent-based tool selection, which uses similarity search over prompt–solution pairs to recommend tools inside an agent, and workflow memory, which uses prompt–solution pairs at the workflow level to recommend which agents to involve. The combination of separation and memory is what distinguishes the system from static single-agent chains.

What would settle it

Take 100 randomly sampled tasks from the 2,000-task benchmark and have remote-sensing analysts write gold tool-call sequences from the raw data products, without seeing GPT-4o's outputs. Score GeoLLM-Squad and the single-agent baseline against those analyst golds; the 17-point gap is supported only if it persists on human-written ground truth. A second check: inspect whether any gold solutions require tool calls that the reference executor cannot run, which would make the benchmark partially unexecutable.

Watch

Extended reading notes

Core claim

GeoLLM-Squad outperforms all evaluated methods on agentic correctness while keeping token cost competitive. The central discovery is that a hybrid orchestration scheme—compositional reasoning to write program-like agent schedules plus iterative ledger-style reassessment on failure, with intent-based tool selection and workflow memory—lets a team of specialized agents outperform both monolithic single-agent copilots and prior multi-agent frameworks. Across five remote sensing workflows and 2,000 tasks, the system reaches 60.29% correct tool-calling sequences, against 39.84–43.32% for the single-agent baseline and 30.08–41.03% for the two multi-agent baselines, at 78.49k tokens per task. Domain-level errors are lowest for NDVI, land surface temperature, and tree loss/canopy, and vision detection F1 is 78.58%. When the orchestrator and agents run on an open-weight 7B model instead of the default serving model, correctness is 40.29% versus 25.54% and 9.29% for the two multi-agent baselines, evidence that the architecture is more tolerant of weaker open models.

Load-bearing premise

The benchmark's ground truth is mostly machine-generated: GPT-4o writes the gold solutions for 2,000 tasks after few-shot training on 56 human-vetted seeds, so the correctness scores are measured against a standard produced by the same model family that runs the agents.

Editorial extensions

If this is right

  • If the results hold, geospatial copilots can scale to multi-domain requests without needing a larger model or longer context window.
  • The 17-point correctness gap suggests that orchestration design matters more than monolithic model scale once toolkits reach hundreds of functions.
  • Open-weight 7B models become usable as agent backends for remote sensing, reaching around 40% correctness instead of collapsing to near-noise levels.
  • Compositional scheduling without frequent scheduler calls keeps token costs near the single-agent baseline while adding multi-agent flexibility.
  • Benchmarks that measure whether the right data subsets were accessed can complement text-similarity or success-rate metrics for agent evaluation.

Reading between the lines

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

  • Editorial inference: because the benchmark's gold solutions are generated by the same model family that runs the agents, part of the 17-point gap may reflect format familiarity rather than task competence; an analyst-written gold set would show how much of the gain is architectural.
  • Editorial inference: the orchestration split could generalize to other tool-heavy scientific domains where a single model's context window is the bottleneck and tools number in the hundreds.
  • Editorial inference: the paper tests functional dependencies (one agent's output feeds another) but not deeper semantic collaboration, so a natural next test is cross-agent reconciliation of conflicting data sources.
  • Editorial inference: the scalability claim rests on combining tasks rather than expanding geographic scope; a stress test with more regions and data volumes would clarify whether the bottleneck is context length or data retrieval.
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 / 7 minor

Summary. The paper presents GeoLLM-Squad, a multi-agent geospatial copilot for remote sensing (RS) workflows that separates an orchestrator from specialized sub-agents, each with its own toolkit. Built on AutoGen and GeoLLM-Engine, the system augments agents with tool selection (TS) and workflow memory (WM) and is evaluated on 2k generated tasks across five RS domains (agriculture, climate, urban, forestry, vision) using GPT-4o-mini as the backend LLM. The authors report a 17% improvement in agentic correctness over the single-agent GeoLLM-Engine baseline (60.29% vs 43.32%) at 78.49k tokens per task, and show that GeoLLM-Squad scales better than single-agent and existing multi-agent baselines as tasks are combined. Additional experiments with Qwen-2.5-7B/3B indicate that GeoLLM-Squad is more robust to open-source small language models than the baselines.

Significance. The paper addresses a relevant and timely problem: scaling LLM-based geospatial copilots to multi-domain remote sensing workflows. The multi-agent design with explicit orchestration, tool selection, and workflow memory is a sensible engineering contribution, and the evaluation covers a broad set of real geospatial products and datasets. However, the central claim rests on an evaluation whose gold labels are generated by GPT-4o from 56 human-vetted seeds, and the same model family is used for the agents and for TS/WM few-shot priming. This creates a circularity risk: correctness scores may measure agreement with GPT-4o's preferred tool-calling order rather than robust performance on human-verified RS workflows. The headline '17% improvement' also conflates percentage points with relative improvement, and no error bars or significance tests are reported. If the evaluation were anchored to a human-verified subset and the metric clarified, the contribution would be substantially stronger; as presented, the central claim is defensible but not yet fully supported.

major comments (3)
  1. [Section III (Task generation) and Section IV (Metrics)] The ground-truth gold solutions for the 2k evaluation tasks are generated by an 'oracle GPT-4o' after few-shot conditioning on 56 human-vetted seeds (Section III, Task generation). The same model family (GPT-4o-mini) is used to run all agents, and the TS and WM components are primed with prompt-solution pairs derived from this generation process. This creates a circularity concern: correctness is measured as agreement with GPT-4o's preferred tool-calling order, so the 60.29% versus 43.32% gap (Table III) may partly reflect GPT-4o-mini's ability to mimic GPT-4o preferences rather than a robust advantage in human-verified RS workflows. I recommend adding a human-annotated evaluation subset (e.g., 100-200 held-out tasks) or a perturbation analysis that shows the ranking is stable to label noise.
  2. [Section IV (Results) and Table III] The headline '17% improvement' conflates percentage points with relative improvement: 60.29% - 43.32% = 16.97 percentage points, which is a 39.2% relative increase over GeoLLM-Engine. The abstract and conclusion state '17% improvement' without specifying which measure is intended. Additionally, no error bars, confidence intervals, or significance tests are reported across repeated runs or task subsets, so it is unclear whether the 16.97-point gap is stable. Please report the relevant metric explicitly and provide variance estimates or significance tests.
  3. [Section III (Agent tools, Orchestrator) and Section IV (Scaling)] The design of GeoLLM-Squad includes 521 API functions (almost 3x the single-agent baseline) and TS/WM few-shot memory sets whose sizes and similarity thresholds are free parameters. The paper does not ablate the contribution of these factors to the correctness gain. Without ablations that vary tool count, memory size, and number of agents, the claimed benefit of 'separating agentic orchestration from geospatial task-solving' is confounded with the increased toolset and few-shot guidance. Please include ablations isolating the orchestration scheme from these other components.
minor comments (7)
  1. [Section IV (Metrics)] The definition of correctness rate ('ratio of correct tool-calling steps') is ambiguous; clarify whether it is exact sequence match, partial credit, and how gold and predicted tool calls are aligned.
  2. [Section III (Task generation)] The number of human annotators, their inter-annotator agreement, and how disagreements in the 56 seed solutions were resolved are not reported; please specify.
  3. [Table III] The row for Magentic uses '– -' for the WM column; use a consistent notation for 'not used' across all rows.
  4. [Section V] The claim 'to our knowledge, the first work to apply multi-agency to RS workflows' is strong; a more careful related-work search (e.g., multi-agent LLM systems in earth science) is needed to support it.
  5. [References] References [6] and [21] contain placeholder arXiv identifiers (2502.XXXXX and 2311.12345); these must be completed.
  6. [Figure 2] Figure 2 is referenced in Section IV but not included in the manuscript version provided; ensure the final version includes the figure with labeled axes, error bars, and a description of how multi-task combinations were constructed.
  7. [Section III] The phrase 'GeoLLM-Squad comprise' should be 'GeoLLM-Squad comprises'.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the headline comparison is an empirical benchmark with shared backends and gold labels; machine-generated gold for the 2k tasks is an external-validity risk rather than a construction-level circularity.

full rationale

The paper's central claim is an empirical comparison, not a derivation. GeoLLM-Squad, GeoLLM-Engine, Chameleon, and Magentic are all evaluated with the same GPT-4o-mini backend and the same gold tool-calling and data-access sets, so the reported 17-point gap is not forced by the definition of the metric. Section III's Task generation paragraph states that only the 56 seed solutions are human-vetted, while the 2k evaluation solutions are produced by few-shot prompting GPT-4o; this is a genuine external-validity weakness and should be disclosed more prominently, but it does not make the relative ranking circular because no method's output is defined as the gold solution and no test label is shown to be inserted into the retrieval corpora. The self-citations to GeoLLM-Engine [5], TS [13], and Geckopt [11] are building blocks and are not invoked as uniqueness theorems or as proof of the main result; the multi-agent architecture is independently compared against external baselines and open-source SLMs. No equation in the paper equates a predicted quantity to a fitted input, so no step meets the bar for a construction-level circularity.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central claim rests on a benchmark whose labels and prompts are generated by GPT-4o, not on parameter fitting; however, no free parameters are fitted to data in the usual sense, and the design choices listed above are hand-selected and affect the reported numbers.

free parameters (2)
  • Number of agents and tool division = 5 domain agents plus orchestrator, Map, Database, DataOps; 521 API functions
    Chosen by the authors rather than learned; the scalability and correctness claims depend on this division.
  • TS/WM few-shot memory set = Precompiled prompt-solution pairs from the generated pseudo-dataset
    Used for intent-based tool selection and workflow memory; selected by the authors and injected into agent prompts, affecting measured correctness.
assumptions (4)
  • domain assumption GPT-4o-generated pseudo-tasks are representative of real remote sensing workflows
    Section III Task generation: 250 tasks per agent are generated by oracle GPT-4o with RAG over human-vetted seeds.
  • domain assumption The gold solutions for the 2k evaluation tasks are correct despite being machine-generated
    Only the 56 seed tasks are human-inspected; the remaining solutions are produced by GPT-4o without per-task human verification.
  • domain assumption Agentic correctness (correct tool-calling steps in expected order) is a valid proxy for task success
    Section IV Metrics defines the metric; downstream geospatial errors are reported as secondary evidence but are also compared against the same gold sets.
  • domain assumption The tool implementations and agent workflows reflect expert geospatial practice
    Section III Agent tools says tools emulate existing RS agents and were tested in-house; no external validation of tool correctness is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Multi-Agent Geospatial Copilots for Remote Sensing Workflows." pith.science (2026). https://pith.science/paper/ZE57MP73

@misc{pith2026250116254,
  author       = {Pith},
  title        = {Pith review of: Multi-Agent Geospatial Copilots for Remote Sensing Workflows},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZE57MP73}},
  note         = {Machine review of arXiv:2501.16254}
}
read the original abstract

We present GeoLLM-Squad, a geospatial Copilot that introduces the novel multi-agent paradigm to remote sensing (RS) workflows. Unlike existing single-agent approaches that rely on monolithic large language models (LLM), GeoLLM-Squad separates agentic orchestration from geospatial task-solving, by delegating RS tasks to specialized sub-agents. Built on the open-source AutoGen and GeoLLM-Engine frameworks, our work enables the modular integration of diverse applications, spanning urban monitoring, forestry protection, climate analysis, and agriculture studies. Our results demonstrate that while single-agent systems struggle to scale with increasing RS task complexity, GeoLLM-Squad maintains robust performance, achieving a 17% improvement in agentic correctness over state-of-the-art baselines. Our findings highlight the potential of multi-agent AI in advancing RS workflows.

Figures

Figures reproduced from arXiv: 2501.16254 by the authors.

Figure 1
Figure 1. GeoLLM-Squad is a multi-agent Copilot for RS workflows that builds on the open-source AutoGen [9] and GeoLLM-Engine [5] frameworks. Our orchestrator supports various geospatial agents and API tools to handle diverse workflows such as climate analysis or urban planning tasks. TABLE II OVERVIEW OF GEOSPATIAL PRODUCTS, METRICS, AND DATASETS USED IN GEOLLM-SQUAD WORKFLOWS. Application Metric Dataset/Product Resolution/D… view at source ↗
Figure 2
Figure 2. Single- vs. multi-agent scalability ablations: we evaluate each method (with TS [13] and WM [14] incorporated) separately for each task and for multi-task combinations. GeoLLM-Squad consistently provides better agentic performance across all tasks. shows Chameleon outperforming GeoEngine-LLM in some tasks, while the single-agent performs better in others. Over￾all, GeoLLM-Squad consistently provides better correctne… view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. CangLing-KnowFlow: A Unified Knowledge-and-Flow-fused Agent for Comprehensive Remote Sensing Applications

    cs.AI 2025-12 reject novelty 5.0 of 10

    CangLing-KnowFlow combines a procedural knowledge base, dynamic workflow repair, and memory to beat ReAct/Reflexion on remote-sensing workflow tasks, but the benchmark is drawn from the same tasks used to build its kn...

Reference graph

Works this paper leans on

61 extracted references · 41 canonical work pages · cited by 1 Pith paper

  1. [1]

    Geochat: Grounded large vision-language model for remote sensing,

    K. Kuckreja, M. S. Danish, M. Naseer, A. Das, S. Khan, and F. S. Khan, “Geochat: Grounded large vision-language model for remote sensing,” in Proceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition (CVPR), June 2024, pp. 27 831–27 840

    work page 2024

  2. [2]

    Skyeyegpt: Unifying remote sensing vision-language tasks via instruction tuning with large language model,

    Y . Zhan, Z. Xiong, and Y . Yuan, “Skyeyegpt: Unifying remote sensing vision-language tasks via instruction tuning with large language model,” arXiv preprint arXiv:2401.09712 , 2024

    arXiv 2024

  3. [3]

    Evaluating tool- augmented agents in remote sensing platforms,

    S. Singh, M. Fore, and D. Stamoulis, “Evaluating tool- augmented agents in remote sensing platforms,” arXiv preprint arXiv:2405.00709, 2024

    arXiv 2024

  4. [4]

    Remote sensing chatgpt: Solving remote sensing tasks with chatgpt and visual models,

    H. Guo, X. Su, C. Wu, B. Du, L. Zhang, and D. Li, “Remote sensing chatgpt: Solving remote sensing tasks with chatgpt and visual models,” in IGARSS 2024 - 2024 IEEE International Geoscience and Remote Sensing Symposium , 2024, pp. 11 474– 11 478

    work page 2024

  5. [5]

    Geollm-engine: A realis- tic environment for building geospatial copilots,

    S. Singh, M. Fore, and D. Stamoulis, “Geollm-engine: A realis- tic environment for building geospatial copilots,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, June 2024, pp. 585–594

    work page 2024

  6. [6]

    Isr-gpt: Evaluating tool-augmented large language models in military isr workflows,

    D. Stamoulis, I. Anagnostopoulos, and C. Hossfeld, “Isr-gpt: Evaluating tool-augmented large language models in military isr workflows,” arXiv preprint arXiv:2502.XXXXX, 2025, under review

    work page 2025

  7. [7]

    Metagpt: Meta programming for multi-agent collaborative framework,

    S. Hong, X. Zheng, J. Chen, Y . Cheng, J. Wang, C. Zhang, Z. Wang, S. K. S. Yau, Z. Lin, L. Zhou et al., “Metagpt: Meta programming for multi-agent collaborative framework,” arXiv preprint arXiv:2308.00352, 2023

    arXiv 2023

  8. [8]

    Magentic-one: A generalist multi-agent system for solving complex tasks,

    A. Fourney, G. Bansal, H. Mozannar, C. Tan, E. Salinas, F. Niedtner, G. Proebsting, G. Bassman, J. Gerrits, J. Alber et al. , “Magentic-one: A generalist multi-agent system for solving complex tasks,” arXiv preprint arXiv:2411.04468, 2024

    arXiv 2024

Show all 61 references
  1. [9]

    Autogen: Enabling next-gen llm applications via multi-agent conversation framework,

    Q. Wu, G. Bansal, J. Zhang, Y . Wu, S. Zhang, E. Zhu, B. Li, L. Jiang, X. Zhang, and C. Wang, “Autogen: Enabling next-gen llm applications via multi-agent conversation framework,”arXiv preprint arXiv:2308.08155, 2023

    2023 arXiv

  2. [10]

    Chameleon: Plug-and-play com- positional reasoning with large language models,

    P. Lu, B. Peng, H. Cheng, M. Galley, K.-W. Chang, Y . N. Wu, S.-C. Zhu, and J. Gao, “Chameleon: Plug-and-play com- positional reasoning with large language models,” Advances in Neural Information Processing Systems , vol. 36, 2024

    2024

  3. [11]

    Geckopt: Llm system efficiency via intent-based tool selection,

    M. Fore, S. Singh, and D. Stamoulis, “Geckopt: Llm system efficiency via intent-based tool selection,” in Proceedings of the Great Lakes Symposium on VLSI 2024 , 2024, pp. 353–354

    2024

  4. [12]

    Webarena: A realistic web environment for building autonomous agents,

    S. Zhou, F. F. Xu, H. Zhu, X. Zhou, R. Lo, A. Sridhar, X. Cheng, T. Ou, Y . Bisk, D. Fried, U. Alon, and G. Neubig, “Webarena: A realistic web environment for building autonomous agents,” in The Twelfth International Conference on Learning Represen- tations, 2024

    2024

  5. [13]

    Less is more: Optimizing function calling for llm execution on edge devices,

    V . Paramanayakam, A. Karatzas, I. Anagnostopoulos, and D. Stamoulis, “Less is more: Optimizing function calling for llm execution on edge devices,” arXiv preprint arXiv:2411.15399 , 2024

    2024 arXiv

  6. [14]

    Agent workflow memory,

    Z. Z. Wang, J. Mao, D. Fried, and G. Neubig, “Agent workflow memory,” arXiv preprint arXiv:2409.07429 , 2024

    2024 arXiv

  7. [15]

    Oreole-fm: successes and challenges toward billion-parameter foundation models for high-resolution satel- lite imagery,

    P. Dias, A. Tsaris, J. Bowman, A. Potnis, J. Arndt, H. L. Yang, and D. Lunga, “Oreole-fm: successes and challenges toward billion-parameter foundation models for high-resolution satel- lite imagery,” in Proceedings of the 32nd ACM International Conference on Advances in Geogra...

    2024

  8. [16]

    Remote sensing vision-language foundation models without annotations via ground remote alignment,

    U. Mall, C. P. Phoo, M. K. Liu, C. V ondrick, B. Hariharan, and K. Bala, “Remote sensing vision-language foundation models without annotations via ground remote alignment,” in The Twelfth International Conference on Learning Representations , 2024

    2024

  9. [17]

    Rsgpt: A remote sensing vision language model and benchmark,

    Y . Hu, J. Yuan, C. Wen, X. Lu, and X. Li, “Rsgpt: A remote sensing vision language model and benchmark,” arXiv preprint arXiv:2307.15266, 2023

    2023 arXiv

  10. [18]

    Large language models for captioning and retrieving remote sensing images,

    J. D. Silva, J. Magalh ˜aes, D. Tuia, and B. Martins, “Large language models for captioning and retrieving remote sensing images,” arXiv preprint arXiv:2402.06475 , 2024

    2024 arXiv

  11. [19]

    Remoteclip: A vision language foundation model for remote sensing,

    F. Liu, D. Chen, Z. Guan, X. Zhou, J. Zhu, Q. Ye, L. Fu, and J. Zhou, “Remoteclip: A vision language foundation model for remote sensing,” IEEE Transactions on Geoscience and Remote Sensing, vol. 62, pp. 1–16, 2024

    2024

  12. [20]

    Earthgpt: A universal multimodal large language model for multisensor image comprehension in remote sensing domain,

    W. Zhang, M. Cai, T. Zhang, Y . Zhuang, and X. Mao, “Earthgpt: A universal multimodal large language model for multisensor image comprehension in remote sensing domain,” IEEE Trans- actions on Geoscience and Remote Sensing , vol. 62, pp. 1–20, 2024

    2024

  13. [21]

    Stable diffusion for aerial object detection,

    Y . Jian, F. Yu, S. Singh, and D. Stamoulis, “Stable diffusion for aerial object detection,” arXiv preprint arXiv:2311.12345 , 2023

    2023 arXiv

  14. [22]

    Urban mobility assessment using llms,

    P. Bhandari, A. Anastasopoulos, and D. Pfoser, “Urban mobility assessment using llms,” in Proceedings of the 32nd ACM Inter- national Conference on Advances in Geographic Information Systems, 2024, pp. 67–79

    2024

  15. [23]

    Harnessing llms for cross-city od flow prediction,

    C. Yu, X. Xie, Y . Huang, and C. Qiu, “Harnessing llms for cross-city od flow prediction,” in Proceedings of the 32nd ACM International Conference on Advances in Geographic Information Systems, 2024, pp. 384–395

    2024

  16. [24]

    Context-aware conversational map search with llm,

    C. Zhang, A. Sriram, K.-H. Hung, R. Wang, and D. Yankov, “Context-aware conversational map search with llm,” in Pro- ceedings of the 32nd ACM International Conference on Ad- vances in Geographic Information Systems , 2024, pp. 485–488

    2024

  17. [25]

    Metafruit meets foundation models: Leveraging a comprehen- sive multi-fruit dataset for advancing agricultural foundation models,

    J. Li, K. Lammers, X. Yin, X. Yin, L. He, R. Lu, and Z. Li, “Metafruit meets foundation models: Leveraging a comprehen- sive multi-fruit dataset for advancing agricultural foundation models,” arXiv preprint arXiv:2407.04711 , 2024

    2024 arXiv

  18. [26]

    Multimodal large language model for wheat breeding: a new exploration of smart breeding,

    G. Yang, Y . Li, Y . He, Z. Zhou, L. Ye, H. Fang, Y . Luo, and X. Feng, “Multimodal large language model for wheat breeding: a new exploration of smart breeding,” arXiv preprint arXiv:2411.15203, 2024

    2024 arXiv

  19. [27]

    Evolving microsoft azure data manager for agriculture to transform data into intuitive insights,

    Microsoft Industry Clouds, “Evolving microsoft azure data manager for agriculture to transform data into intuitive insights,” https://azure.microsoft.com/en-us/blog/ evolving-microsoft-azure-data-manager-for-agriculture, November 2023, accessed: 2025-01-07

    2023

  20. [28]

    Fomo-bench: a multi-modal, multi-scale and multi-task forest monitoring benchmark for remote sensing foundation models,

    N. I. Bountos, A. Ouaknine, and D. Rolnick, “Fomo-bench: a multi-modal, multi-scale and multi-task forest monitoring benchmark for remote sensing foundation models,” arXiv preprint arXiv:2312.10114, 2023

    2023 arXiv

  21. [29]

    Geo-bench: Toward foundation models for earth monitoring,

    A. Lacoste, N. Lehmann, P. Rodriguez, E. Sherwin, H. Kerner, B. L¨utjens, J. Irvin, D. Dao, H. Alemohammad, A. Drouin et al., “Geo-bench: Toward foundation models for earth monitoring,” Advances in Neural Information Processing Systems , vol. 36, 2024

    2024

  22. [30]

    On the foundations of earth and climate foundation models,

    X. X. Zhu, Z. Xiong, Y . Wang, A. J. Stewart, K. Heidler, Y . Wang, Z. Yuan, T. Dujardin, Q. Xu, and Y . Shi, “On the foundations of earth and climate foundation models,” arXiv preprint arXiv:2405.04285, 2024

    2024 arXiv

  23. [31]

    Neu- ral plasticity-inspired foundation model for observing the earth crossing modalities,

    Z. Xiong, Y . Wang, F. Zhang, A. J. Stewart, J. Hanna, D. Borth, I. Papoutsis, B. Le Saux, G. Camps-Valls, and X. X. Zhu, “Neu- ral plasticity-inspired foundation model for observing the earth crossing modalities,” arXiv e-prints, pp. arXiv–2403, 2024

    2024

  24. [32]

    From data to decisions: Streamlining geospa- tial operations with multimodal globeflowgpt,

    D. Kononykhin, M. Mozikov, K. Mishtal, P. Kuznetsov, D. Abramov, N. Sotiriadi, Y . Maximov, A. V . Savchenko, and I. Makarov, “From data to decisions: Streamlining geospa- tial operations with multimodal globeflowgpt,” in Proceedings of the 32nd ACM International Conference on...

    2024

  25. [33]

    An llm agent for automatic geospatial data analysis,

    Y . Chen, W. Wang, S. Lobry, and C. Kurtz, “An llm agent for automatic geospatial data analysis,” arXiv preprint arXiv:2410.18792, 2024

    2024 arXiv

  26. [34]

    Disasterresponsegpt: Large language models for accelerated plan of action de- velopment in disaster response scenarios,

    V . G. Goecks and N. R. Waytowich, “Disasterresponsegpt: Large language models for accelerated plan of action de- velopment in disaster response scenarios,” arXiv preprint arXiv:2306.17271, 2023

    2023 arXiv

  27. [35]

    Mod13a3 modis/terra vegetation indices monthly l3 global 1km sin grid v006,

    K. Didan, “Mod13a3 modis/terra vegetation indices monthly l3 global 1km sin grid v006,” 2015, accessed 2025- 01-07. [Online]. Available: https://doi.org/10.5067/MODIS/ MOD13A3.006

    2015 doi

  28. [36]

    Mod09ga modis/terra surface reflectance daily l2g global 1km sin grid v006,

    E. Vermote, C. Justice, M. Claverie, and B. Franch, “Mod09ga modis/terra surface reflectance daily l2g global 1km sin grid v006,” 2015, accessed 2025-01-07. [Online]. Available: https://doi.org/10.5067/MODIS/MOD09GA.006

    2015 doi

  29. [37]

    Modis/terra land surface temperature/emissivity 8-day l3 global 1km sin grid v061,

    Z. Wan, S. Hook, and G. Hulley, “Modis/terra land surface temperature/emissivity 8-day l3 global 1km sin grid v061,” https://doi.org/10.5067/MODIS/MYD11A2.061, 2021, accessed 2025-01-07

    2021 doi

  30. [38]

    MCD19A2 MODIS/Terra+Aqua Land Aerosol Optical Depth Daily L2G Global 1km SIN Grid V006,

    A. Lyapustin and Y . Wang, “MCD19A2 MODIS/Terra+Aqua Land Aerosol Optical Depth Daily L2G Global 1km SIN Grid V006,” https://doi.org/10.5067/MODIS/MCD19A2.006, 2018, accessed 2025-01-07

    2018 doi

  31. [39]

    Commission and J

    E. Commission and J. R. Centre, GHSL data package 2023 . Publications Office of the European Union, 2023

    2023

  32. [40]

    Pesaresi and P

    M. Pesaresi and P. Politis, GHS-BUILT-S R2023A - GHS built-up surface grid, derived from Sentinel2 composite and Landsat, multitemporal (1975-2030) . European Commission, Joint Research Centre (JRC), 2023. [Online]. Available: http: //data.europa.eu/89h/9f06f36f-4b11-47ec-abb0...

    1975

  33. [41]

    High-resolution global maps of 21st- century forest cover change,

    M. C. Hansen, P. V . Potapov, R. Moore, M. Hancher, S. A. Turubanova, A. Tyukavina, D. Thau, S. V . Stehman, S. J. Goetz, T. R. Loveland et al. , “High-resolution global maps of 21st- century forest cover change,” science, vol. 342, no. 6160, pp. 850–853, 2013

    2013

  34. [42]

    xview: Objects in context in overhead imagery,

    D. Lam, R. Kuzma, K. McGee, S. Dooley, M. Laielli, M. Klaric, Y . Bulatov, and B. McCord, “xview: Objects in context in overhead imagery,” arXiv preprint arXiv:1802.07856 , 2018

    2018 arXiv

  35. [43]

    Creating xbd: A dataset for assessing building damage from satellite imagery,

    R. Gupta, B. Goodman, N. Patel, R. Hosfelt, S. Sajeev, E. Heim, J. Doshi, K. Lucas, H. Choset, and M. Gaston, “Creating xbd: A dataset for assessing building damage from satellite imagery,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition wo...

    2019

  36. [44]

    xview3-sar: Detect- ing dark fishing activity using synthetic aperture radar imagery,

    F. Paolo, T.-t. T. Lin, R. Gupta, B. Goodman, N. Patel, D. Kuster, D. Kroodsma, and J. Dunnmon, “xview3-sar: Detect- ing dark fishing activity using synthetic aperture radar imagery,” Advances in Neural Information Processing Systems , vol. 35, pp. 37 604–37 616, 2022

    2022

  37. [45]

    The sarfish dataset and challenge,

    C. Luckett, B. McCarthy, T.-T. Cao, and A. Robles-Kelly, “The sarfish dataset and challenge,” in Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision , 2024, pp. 752–761

    2024

  38. [46]

    Fair1m: A benchmark dataset for fine-grained object recognition in high-resolution remote sens- ing imagery,

    X. Sun, P. Wang, Z. Yan, F. Xu, R. Wang, W. Diao, J. Chen, J. Li, Y . Feng, T. Xu et al., “Fair1m: A benchmark dataset for fine-grained object recognition in high-resolution remote sens- ing imagery,” ISPRS Journal of Photogrammetry and Remote Sensing, vol. 184, pp. 116–130, 2022

    2022

  39. [47]

    Func- tional map of the world,

    G. Christie, N. Fendley, J. Wilson, and R. Mukherjee, “Func- tional map of the world,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition , 2018, pp. 6172– 6180

    2018

  40. [48]

    Bigearth- net: A large-scale benchmark archive for remote sensing im- age understanding,

    G. Sumbul, M. Charfuelan, B. Demir, and V . Markl, “Bigearth- net: A large-scale benchmark archive for remote sensing im- age understanding,” in IGARSS 2019-2019 IEEE International Geoscience and Remote Sensing Symposium . IEEE, 2019, pp. 5901–5904

    2019

  41. [49]

    NASA Earthdata Search,

    Earth Science Data and Information System (ESDIS), “NASA Earthdata Search,” https://search.earthdata.nasa.gov, 2025, ac- cessed 2025-01-07

    2025

  42. [50]

    Gpt-4o-mini: Advancing cost- efficient intelligence,

    OpenAI, “Gpt-4o-mini: Advancing cost- efficient intelligence,” https://openai.com/index/ gpt-4o-mini-advancing-cost-efficient-intelligence/, 2025, accessed: January 15, 2025

    2025

  43. [51]

    Gpt-4o system card,

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

    2024 arXiv

  44. [52]

    Toolqa: A dataset for llm question answering with external tools,

    Y . Zhuang, Y . Yu, K. Wang, H. Sun, and C. Zhang, “Toolqa: A dataset for llm question answering with external tools,” Advances in Neural Information Processing Systems , vol. 36, pp. 50 117–50 143, 2023

    2023

  45. [53]

    Cook- book: A framework for improving llm generative abilities via programmatic data generating templates,

    A. Narayan, M. F. Chen, K. Bhatia, and C. R ´e, “Cook- book: A framework for improving llm generative abilities via programmatic data generating templates,” arXiv preprint arXiv:2410.05224, 2024

    2024 arXiv

  46. [54]

    Nexusraven: a commercially- permissive language model for function calling,

    V . K. Srinivasan, Z. Dong, B. Zhu, B. Yu, D. Mosk-Aoyama, K. Keutzer, J. Jiao, and J. Zhang, “Nexusraven: a commercially- permissive language model for function calling,” in NeurIPS 2023 Foundation Models for Decision Making Workshop, 2023

    2023

  47. [55]

    Visualwebarena: Evaluating multimodal agents on realistic visual web tasks,

    J. Y . Koh, R. Lo, L. Jang, V . Duvvur, M. C. Lim, P.-Y . Huang, G. Neubig, S. Zhou, R. Salakhutdinov, and D. Fried, “Visualwebarena: Evaluating multimodal agents on realistic visual web tasks,” in ICLR 2024 Workshop on Large Language Model (LLM) Agents , 2024

    2024

  48. [56]

    Neu- ralpower: Predict and deploy energy-efficient convolutional neural networks,

    E. Cai, D.-C. Juan, D. Stamoulis, and D. Marculescu, “Neu- ralpower: Predict and deploy energy-efficient convolutional neural networks,” in Asian Conference on Machine Learning . PMLR, 2017, pp. 622–637

    2017

  49. [57]

    Qwen2. 5 technical report,

    A. Yang, B. Yang, B. Zhang, B. Hui, B. Zheng, B. Yu, C. Li, D. Liu, F. Huang, H. Wei et al. , “Qwen2. 5 technical report,” arXiv preprint arXiv:2412.15115 , 2024

    2024 arXiv

  50. [58]

    vllm documentation,

    vLLM Team, “vllm documentation,” https://docs.vllm.ai/en/ stable/, 2025, accessed: January 15, 2025

    2025

  51. [59]

    Data-centric machine learn- ing for geospatial remote sensing data,

    R. Roscher, M. Rußwurm, C. Gevaert, M. Kampffmeyer, J. A. dos Santos, M. Vakalopoulou, R. H ¨ansch, S. Hansen, K. Nogueira, J. Prexl et al. , “Data-centric machine learn- ing for geospatial remote sensing data,” arXiv preprint arXiv:2312.05327, 2023

    2023 arXiv

  52. [60]

    Eval- uating digital tools for sustainable agriculture using causal inference,

    I. Tsoumas, G. Giannarakis, V . Sitokonstantinou, A. Koukos, D. Loka, N. Bartsotas, C. Kontoes, and I. Athanasiadis, “Eval- uating digital tools for sustainable agriculture using causal inference,” arXiv preprint arXiv:2211.03195 , 2022

    2022 arXiv

  53. [61]

    Global prevalence of setting long- lines at dawn highlights bycatch risk for threatened albatross,

    D. Kroodsma, J. Turner, C. Luck, T. Hochberg, N. Miller, P. Augustyn, and S. Prince, “Global prevalence of setting long- lines at dawn highlights bycatch risk for threatened albatross,” Biological Conservation, vol. 283, p. 110026, 2023

    2023

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.