OpenAg: Democratizing Agricultural Intelligence

Jason Fisher; Srikanth Thudumu

REVIEW 3 major objections 5 minor 25 references

OpenAg: Democratizing Agricultural Intelligence

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read OpenAg proposes a six-part AI architecture that would make agricultural decision support contextual, explainable, and adaptive.

desk verdict A solid, readable architecture blueprint for agricultural AI, but the central tacit-knowledge bridge is untested and the capability claims run ahead of any implementation. read the letter →

arxiv 2506.04571 v2 pith:N6EXHRK7 submitted 2025-06-05 cs.AI

Srikanth Thudumu , Jason Fisher This is my paper

classification cs.AI

keywords AgriculturalAGIdecisionsupportNeuralknowledgegraphsMulti-agentreasoningCausalexplainabilityTransferlearningTacitintegrationSmallholderfarming

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

OpenAg is a proposed architecture for agricultural intelligence that weaves together structured knowledge ingestion, a neural knowledge graph, specialist multi-agent reasoning, causal explanations, and transfer learning into a single pipeline. The paper argues that no existing approach—general-purpose large language models, static knowledge graphs, or black-box machine learning—can alone give farmers the contextual understanding, explainability, and adaptability that real agricultural decisions require, especially for smallholder farmers with limited resources. The central claim is that integrating these components in one framework delivers context-aware, explainable, and actionable recommendations and bridges scientific knowledge with the tacit expertise of experienced farmers. A sympathetic reader would care because the architecture directly targets the adoption barriers that have kept prior agricultural decision-support systems out of practical use.

What carries the argument

The load-bearing object is the OpenAg architecture itself, with the Neural Agricultural Knowledge Graph Engine as the central integrative mechanism. That component converts the UAKB's structured facts into a learnable graph representation using Graph Neural Networks and knowledge graph embeddings, enabling relational, temporal, and spatial inference. From there, the Adaptive Multi-Agent Reasoning System coordinates specialist agents, and the Causal Agricultural Decision Transparency mechanism translates the reasoning into causal, counterfactual, and uncertainty-aware explanations. The claim is that this flow—not any single technique in isolation—is what makes agricultural intelligence contextual, explainable, and adaptive.

What would settle it

Conduct a controlled comparison in which the UAKB is built from scientific literature alone, then augmented with knowledge elicited from experienced farmers; if recommendation quality on smallholder plots fails to improve measurably after the farmer knowledge is added, the bridging claim is refuted.

Watch

Extended reading notes

Core claim

The paper's central claim is a systems-level one: agricultural AI falls short in practice because knowledge representation, reasoning, explanation, and adaptation are developed separately, so each component fails in the field. OpenAg is designed so that these functions reinforce each other across six components: multi-modal knowledge ingestion, the Unified Agricultural Knowledge Base (UAKB), the Neural Agricultural Knowledge Graph Engine, the Adaptive Multi-Agent Reasoning System (AMRS), the Causal Agricultural Decision Transparency mechanism (CADET), and Adaptive Agricultural Transfer Learning. The neural knowledge graph turns static facts into learnable representations; specialist agents collaborate through consensus protocols to produce recommendations; causal inference and counterfactual reasoning explain those recommendations; and transfer learning lets the system adapt to new regions with minimal data. The paper's intended contribution is to show how these pieces fit together into an architecture that can move toward agricultural general intelligence.

Load-bearing premise

The framework's central promise depends on the assumption that a farmer's tacit expertise can be elicited through structured interviews and knowledge engineering and encoded in the Unified Agricultural Knowledge Base without losing the context that makes it useful.

Editorial extensions

If this is right

If OpenAg works as specified, farmers would receive recommendations whose causal rationales, counterfactual alternatives, and confidence levels are made explicit, lowering the trust barrier that has blocked adoption of decision-support systems.
If the transfer-learning component performs as proposed, OpenAg could be deployed in data-scarce regions with only a few local examples, reducing the need for region-specific training data.
If farmer tacit knowledge can be encoded in the UAKB, recommendations would combine peer-reviewed literature with localized experiential knowledge, making advice relevant to smallholdings rather than generic.
If the neural knowledge graph generalizes, OpenAg could discover relationships not explicitly stated in any source document, generating new hypotheses about soil, crop, pest, and climate interactions.
If multi-agent consensus operates as designed, the system would handle decisions that require simultaneous expertise in crops, resources, risk, and sustainability in a single coherent recommendation.

Reading between the lines

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

I infer that the tacit-knowledge handling is the part that most distinguishes this architecture from an LLM-based agricultural chatbot; a testable consequence is that the quality of farmer interviews, not model size, will be the binding constraint on recommendation quality.
A natural extension the paper leaves implicit is an ablation protocol: running the multi-agent system with and without the neural knowledge graph would isolate which component actually improves explainability or accuracy.
The causal claims are checkable on longitudinal farm records; for example, a counterfactual such as 'what would yield have been without the recommended intervention' can be computed once outcome data accumulate.
Because the paper is architectural rather than empirical, the next concrete step is a reference implementation on one crop and region; only then can the claimed integration benefits be tested.

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.

Desk Editor's Note

read the letter

Read OpenAg. It's a blueprint, not a built system, and the stress-test is right: the paper's central promise—bridging scientific knowledge and farmer tacit knowledge—is also its weakest link. Section 1 says tacit knowledge is 'not easily articulated'; Section 4.1 says you capture it with structured interviews. Those two claims sit uneasily together, and the paper never acknowledges the tension or proposes a validation check.

What's new and useful: the paper names real adoption bottlenecks (black-box recommendations, poor local adaptation) and assembles a more complete pipeline than most agri-AI proposals—UAKB, neural knowledge graph, multi-agent reasoning, causal explainability, transfer learning. As a design skeleton for what a full decision-support system would need, it's coherent and readable. The literature grounding is honest; the citations are background context, including the single self-citation.

Soft spots, in proportion. First, the capability claims overreach. The abstract says the system 'delivers' context-aware insights; Section 3.5 says CADET 'ensures' interpretability. With no implementation, dataset, or evaluation, those verbs should be 'aims to'. That's a framing issue, not a fatal one, but it invites readers to think the system exists. Second, the tacit-knowledge bridge is load-bearing. Every other component feeds the UAKB; if structured interviews can't capture the context-dependence (soil feel, seasonal timing, risk heuristics), the system is just scientific literature plus questionnaire responses. The paper gives no elicitation protocol, no representation scheme for non-propositional expertise, and no test that the encoded knowledge reproduces farmer decisions. That's a genuine design gap, not merely a missing experiment. Third, the 'agricultural AGI' label is generous for an unbuilt proposal. Minor: the RotatE citation is misspelled 'Rotate'—small sloppiness.

Who it's for: a reader mapping the agri-AI design space, or a group scoping an actual system, could use this as a checklist. A reader wanting evidence that any component works will not find it.

Recommendation: I'd send it to peer review, with a referee tasked to press on the tacit-knowledge elicitation and demand that capability claims be downgraded to aspirations. The problem is real, the architecture is a coherent starting point, and the conversation is worth having. My own verdict is skeptical as written, but this deserves serious engagement rather than a desk reject.

Referee Report

3 major / 5 minor

Summary. The manuscript proposes OpenAg, a six-component architecture for agricultural decision support: multi-modal knowledge ingestion, a Unified Agricultural Knowledge Base (UAKB), a Neural Agricultural Knowledge Graph Engine, an Adaptive Multi-Agent Reasoning System (AMRS), a Causal Agricultural Decision Transparency (CADET) mechanism, and an Adaptive Agricultural Transfer Learning component. The abstract and conclusion claim that this framework will 'deliver context-aware, explainable, and actionable insights' and 'bridge the gap between scientific knowledge and the tacit expertise of experienced farmers.' The paper contains no implementation, no experiments, no datasets, and no formal proofs; it is an architectural blueprint that describes each component at a conceptual level and cites external methods that could be used to build them.

Significance. The paper identifies real, well-documented limitations of current agricultural AI systems: lack of contextualization, black-box recommendations, and poor transfer across regions and crops. Its proposed modular design sensibly assembles existing techniques from knowledge graphs, multi-agent systems, causal inference, and transfer learning into a single vision. If the system were actually built and validated, the architecture could serve as a useful integration template. However, as written, the contribution is a synthesis and a proposal, not a demonstrated system. There is no code, data, machine-checked proof, or falsifiable prediction to anchor the central claims, so the scientific value is conditional on future implementation. The paper does provide a fairly complete enumeration of relevant prior work and component responsibilities, which could be a useful roadmap for practitioners.

major comments (3)

[Abstract; Sections 3.5, 5] The paper repeatedly asserts realized capabilities rather than design intentions. The Abstract says OpenAg 'delivers context-aware, explainable, and actionable insights' and that CADET 'ensures AI recommendations are interpretable'; Section 3.5 says the component 'ensures that all recommendations from OpenAg are transparent'; Section 5 states that 'the proposed architecture transforms heterogeneous fragmented agricultural knowledge into structured, machine-readable intelligence.' None of these assertions is supported by an implementation, a case study, a user study, or a formal analysis, because the manuscript contains only conceptual component descriptions and references to external methods. This is load-bearing because the central claim of the paper is the demonstrated ability to deliver such insights. To make the claim defensible, the authors should either present at least a prototype evaluation or systematically rephrase all capability claims as goals (e.g., 'is designed to...') and add an explicit limitations section.
[Sections 1, 3.2, 4.1] There is an unresolved tension between the paper's definition of tacit knowledge and its proposed capture method. Section 1 defines tacit knowledge as 'intuitive expertise developed through experience that is not easily articulated,' but Section 4.1 proposes to 'capture tacit knowledge through structured interviews and knowledge engineering techniques.' A structured interview that treats knowledge as propositional content cannot, on the paper's own definition, capture the non-articulated, context-dependent component that makes tacit knowledge useful in farming (e.g., soil-feel judgments, seasonal timing heuristics, risk attitudes). The paper provides no elicitation protocol, no representation scheme for non-propositional expertise, and no validation that the knowledge extracted from farmers reproduces the farmers' own decisions. Since the abstract's central promise is to 'bridge the gap between scientific knowledge and the tacit expertise of experienced farmers,' this unresolved issue is load-bearing. A concrete remedy would be to specify a validation protocol, such as comparing system recommendations against farmer decisions on held-out scenarios, or to use observational learning from farmer behavior rather than interviews alone.
[Sections 3.6, 4.5] The Adaptive Agricultural Transfer Learning component is claimed to enable OpenAg 'to adapt its knowledge and recommendations to new agricultural contexts with minimal additional training' (Section 3.6), but the manuscript provides no problem formulation for the domain shift, no definition of 'minimal,' and no evaluation. Section 4.5 lists generic techniques (domain-adversarial training, MAML, elastic weight consolidation, experience replay, knowledge distillation, modular architectures) without specifying which components of OpenAg are transferred, what the source and target domains are, or how adaptation would be measured. Because adaptability is one of the three core limitations the paper claims to solve (Introduction), this is not a minor omission. The authors should either present a concrete transfer scenario with evaluation on at least one crop/region pair, or explicitly label this component as untested future work.

minor comments (5)

[Section 2.1] The sentence 'Sustainable agriculture requires a more sophisticated approach to AI [1]that can overcome...' has a missing space between the citation and 'that.'
[References [6] and [18]] Reference [6] should read 'H. Farhan' rather than 'F ARHAN', and reference [18] should read 'RotatE' rather than 'Rotate' to match the method name used in the text.
[Figure 1] The text refers to Figure 1 and it has a caption, but the figure image is not included in the manuscript text I received; if this is a rendering issue, please ignore, but otherwise the figure needs to be added.
[Section 3.2] The statement 'This component draws inspiration from previous work on knowledge graphs for agricultural applications' has no citation; please add the relevant references so readers can locate the prior work.
[Keywords] The keyword 'Small Language Models' appears in the keyword list, but the paper only discusses large language models; either add a discussion of small/lightweight models or remove the keyword to match the content.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: the paper is an architecture proposal with no fitted parameters, no predictions, and no load-bearing self-citation.

full rationale

OpenAg is a conceptual system description rather than a derivation chain. It contains no equations, fitted parameters, benchmark predictions, or formal theorems whose conclusions could reduce to their inputs. The architecture combines externally cited techniques (knowledge graphs, GNNs, multi-agent consensus, causal inference, transfer learning) and describes how they would be assembled; none of these components is defined in terms of the paper's own claimed outcomes. The single self-citation, reference [23] (Du, Thudumu, Vasa, and Mouzakis), is used in Section 4.3 only as one of three supporting references for a generic consensus-protocol step list, and it is not load-bearing for the central claim. The tacit-knowledge elicitation concern raised by reviewers is a feasibility and validation gap, not a circularity: the paper does not claim to have derived farmer tacit knowledge from the UAKB or vice versa. Because there is no derivation, prediction, or fitted input to be circular about, the appropriate score is 0.

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

The central architecture rests on three feasibility assumptions that are asserted in Sections 3 and 4 but not demonstrated: tacit knowledge can be codified, causal discovery works on agricultural observational data, and multi-agent consensus improves decisions. These are domain assumptions rather than standard math, and none is supported by an experiment in the paper.

assumptions (3)

domain assumption Agricultural knowledge, including tacit farmer expertise, can be elicited and represented in a machine-readable ontology and knowledge graph.
Sections 3.1 and 4.1 describe expert knowledge elicitation via structured interviews and knowledge engineering; the entire Unified Agricultural Knowledge Base design depends on this feasibility.
domain assumption Causal discovery from heterogeneous observational agricultural data can recover reliable cause-effect relationships.
Section 4.4 applies PC, FCI, do-calculus, and related methods to agricultural data, but no evidence is given that these yield valid causal models in this setting.
domain assumption Multi-agent consensus and meta-reasoning improve decision quality over monolithic models.
Sections 3.4 and 4.3 assume that collaboration among specialist agents improves recommendations, but no comparison or evaluation is provided.

how reviews work

0 comments

Cite this review

Pith. "Pith review of OpenAg: Democratizing Agricultural Intelligence." pith.science (2026). https://pith.science/paper/N6EXHRK7

@misc{pith2026250604571,
  author       = {Pith},
  title        = {Pith review of: OpenAg: Democratizing Agricultural Intelligence},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/N6EXHRK7}},
  note         = {Machine review of arXiv:2506.04571}
}

read the original abstract

Agriculture is undergoing a major transformation driven by artificial intelligence (AI), machine learning, and knowledge representation technologies. However, current agricultural intelligence systems often lack contextual understanding, explainability, and adaptability, especially for smallholder farmers with limited resources. General-purpose large language models (LLMs), while powerful, typically lack the domain-specific knowledge and contextual reasoning needed for practical decision support in farming. They tend to produce recommendations that are too generic or unrealistic for real-world applications. To address these challenges, we present OpenAg, a comprehensive framework designed to advance agricultural artificial general intelligence (AGI). OpenAg combines domain-specific foundation models, neural knowledge graphs, multi-agent reasoning, causal explainability, and adaptive transfer learning to deliver context-aware, explainable, and actionable insights. The system includes: (i) a unified agricultural knowledge base that integrates scientific literature, sensor data, and farmer-generated knowledge; (ii) a neural agricultural knowledge graph for structured reasoning and inference; (iii) an adaptive multi-agent reasoning system where AI agents specialize and collaborate across agricultural domains; and (iv) a causal transparency mechanism that ensures AI recommendations are interpretable, scientifically grounded, and aligned with real-world constraints. OpenAg aims to bridge the gap between scientific knowledge and the tacit expertise of experienced farmers to support scalable and locally relevant agricultural decision-making.

Figures

Figures reproduced from arXiv: 2506.04571 by the authors.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

25 extracted references · 21 canonical work pages

[1]

Sustainable ai- based production agriculture: Exploring ai applications and implications in agricultural practices,

A. Mana, A. Allouhi, A. Hamrani, S. Rehman, I. El Jamaoui, and K. Jayachandran, “Sustainable ai- based production agriculture: Exploring ai applications and implications in agricultural practices,” Smart Agricultural Technology, vol. 7, p. 100416, 2024

work page 2024
[2]

Ai-powered decision support systems for sustainable agriculture using ai-chatbot solution,

E. Asolo, I. Gil-Ozoudeh, C. Ejimuda et al. , “Ai-powered decision support systems for sustainable agriculture using ai-chatbot solution,” Journal of Digital Food, Energy & Water Systems , vol. 5, no. 1, 2024

work page 2024
[3]

Oada technical reference: Building interoperable agricultural data sys- tems,

Open Ag Data Alliance, “Oada technical reference: Building interoperable agricultural data sys- tems,” OpenAg Foundation, Technical Report, 2023

work page 2023
[4]

Improving decision support systems with machine learning: Identifying barriers to adoption,

S. Brugler, M. Gardezi, A. Dadkhah, D. M. Rizzo, A. Zia, and S. A. Clay, “Improving decision support systems with machine learning: Identifying barriers to adoption,” Agronomy Journal, vol. 116, no. 3, pp. 1229–1236, 2024

work page 2024
[5]

Deep learning for image-based cassava disease detection,

A. Ramcharan, K. Baranowski, P. McCloskey, B. Ahmed, J. Legg, and D. P. Hughes, “Deep learning for image-based cassava disease detection,” Frontiers in plant science , vol. 8, p. 1852, 2017

work page 2017
[6]

Transfer learning for crop type classification: The agricultural breakthrough with ai,

H. F ARHAN, L. Di, and C. Zhang, “Transfer learning for crop type classification: The agricultural breakthrough with ai,” Journal of Student-Scientists’ Research , vol. 5, 2023

work page 2023
[7]

A comprehensive review of the data and knowledge graphs approaches in bioinfor- matics,

Y. Galluzzo, “A comprehensive review of the data and knowledge graphs approaches in bioinfor- matics,” Computer Science and Information Systems , no. 00, pp. 27–27, 2024

work page 2024
[8]

Knowledge graphs: Opportunities and chal- lenges,

C. Peng, F. Xia, M. Naseriparsa, and F. Osborne, “Knowledge graphs: Opportunities and chal- lenges,” Artificial Intelligence Review , vol. 56, no. 11, pp. 13 071–13 102, 2023

work page 2023

Show all 25 references

[9]

Towards improving prediction accuracy and user-level explainability using deep learning and knowledge graphs: A study on cassava disease,

T. R. Chhetri, A. Hohenegger, A. Fensel, M. A. Kasali, and A. A. Adekunle, “Towards improving prediction accuracy and user-level explainability using deep learning and knowledge graphs: A study on cassava disease,” Expert Systems with Applications , vol. 233, p. 120955, 2023

2023
[10]

Responsible ai in farming: A multi-criteria framework for sus- tainable technology design,

K. Mallinger and R. Baeza-Yates, “Responsible ai in farming: A multi-criteria framework for sus- tainable technology design,” Applied Sciences, vol. 14, no. 1, p. 437, 2024

2024
[11]

Responsible artificial intelli- gence in agriculture requires systemic understanding of risks and externalities,

A. Tzachor, M. Devare, B. King, S. Avin, and S. ´O h ´Eigeartaigh, “Responsible artificial intelli- gence in agriculture requires systemic understanding of risks and externalities,” Nature Machine Intelligence, vol. 4, no. 2, pp. 104–109, 2022

2022
[12]

Enhancing crop recommendation systems with explainable artificial intelligence: a study on agricultural decision-making,

M. Y. Shams, S. A. Gamel, and F. M. Talaat, “Enhancing crop recommendation systems with explainable artificial intelligence: a study on agricultural decision-making,” Neural Computing and Applications, vol. 36, no. 11, pp. 5695–5714, 2024

2024
[13]

Causal reasoning and large language models: Opening a new frontier for causality,

E. Kiciman, R. Ness, A. Sharma, and C. Tan, “Causal reasoning and large language models: Opening a new frontier for causality,” Transactions on Machine Learning Research, 2023

2023
[14]

Building multi-agent copilot towards autonomous agricultural data management and analysis,

Y. Pan, J. Sun, H. Yu, J. Luck, G. Bai, N. Chamara, Y. Ge, and T. Awada, “Building multi-agent copilot towards autonomous agricultural data management and analysis,” in 2024 IEEE Interna- tional Conference on Big Data (BigData) . IEEE, 2024, pp. 4384–4393

2024
[15]

Graph neural networks,

G. Corso, H. Stark, S. Jegelka, T. Jaakkola, and R. Barzilay, “Graph neural networks,” Nature Reviews Methods Primers , vol. 4, no. 1, p. 17, 2024

2024
[16]

Temporal graph networks for deep learning on dynamic graphs,

E. Rossi, B. Chamberlain, F. Frasca, D. Eynard, F. Monti, and M. Bronstein, “Temporal graph networks for deep learning on dynamic graphs,” arXiv preprint arXiv:2006.10637 , 2020

2006 arXiv
[17]

Transe-mtp: A new representation learning method for knowledge graph em- bedding with multi-translation principles and transe,

Y. Li and C. Zhu, “Transe-mtp: A new representation learning method for knowledge graph em- bedding with multi-translation principles and transe,” Electronics, vol. 13, no. 16, p. 3171, 2024

2024
[18]

Rotate: Knowledge graph embedding by relational rotation in complex space,

Z. Sun, Z.-H. Deng, J.-Y. Nie, and J. Tang, “Rotate: Knowledge graph embedding by relational rotation in complex space,” arXiv preprint arXiv:1902.10197 , 2019. 9

1902 arXiv
[19]

Knowledge graph embedding using graph convolutional networks with relation-aware attention,

N. Sheikh, X. Qin, B. Reinwald, C. Miksovic, T. Gschwind, and P. Scotton, “Knowledge graph embedding using graph convolutional networks with relation-aware attention,” arXiv preprint arXiv:2102.07200, 2021

2021 arXiv
[20]

A cooperative multi-agent reinforcement learning framework for resource balancing in complex logistics network,

X. Li, J. Zhang, J. Bian, Y. Tong, and T.-Y. Liu, “A cooperative multi-agent reinforcement learning framework for resource balancing in complex logistics network,” arXiv preprint arXiv:1903.00714 , 2019

1903 arXiv
[21]

Consensus in multi-agent systems: a review,

A. Amirkhani and A. H. Barshooi, “Consensus in multi-agent systems: a review,” Artificial Intelli- gence Review, vol. 55, no. 5, pp. 3897–3935, 2022

2022
[22]

Recent advances in consensus of multi-agent systems: A brief survey,

J. Qin, Q. Ma, Y. Shi, and L. Wang, “Recent advances in consensus of multi-agent systems: A brief survey,” IEEE Transactions on Industrial Electronics , vol. 64, no. 6, pp. 4972–4983, 2016

2016
[23]

A survey on context-aware multi-agent systems: techniques, challenges and future directions,

H. Du, S. Thudumu, R. Vasa, and K. Mouzakis, “A survey on context-aware multi-agent systems: techniques, challenges and future directions,” arXiv preprint arXiv:2402.01968 , 2024

2024 arXiv
[24]

The role of causality in explainable artificial intelligence,

G. Carloni, A. Berti, and S. Colantonio, “The role of causality in explainable artificial intelligence,” Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery , vol. 15, no. 2, p. e70015, 2025

2025
[25]

Reminder of the first paper on transfer learning in neural networks, 1976,

S. Bozinovski, “Reminder of the first paper on transfer learning in neural networks, 1976,” Infor- matica, vol. 44, no. 3, 2020. 10

1976

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

Sustainable ai- based production agriculture: Exploring ai applications and implications in agricultural practices,

A. Mana, A. Allouhi, A. Hamrani, S. Rehman, I. El Jamaoui, and K. Jayachandran, “Sustainable ai- based production agriculture: Exploring ai applications and implications in agricultural practices,” Smart Agricultural Technology, vol. 7, p. 100416, 2024

work page 2024

[2] [2]

Ai-powered decision support systems for sustainable agriculture using ai-chatbot solution,

E. Asolo, I. Gil-Ozoudeh, C. Ejimuda et al. , “Ai-powered decision support systems for sustainable agriculture using ai-chatbot solution,” Journal of Digital Food, Energy & Water Systems , vol. 5, no. 1, 2024

work page 2024

[3] [3]

Oada technical reference: Building interoperable agricultural data sys- tems,

Open Ag Data Alliance, “Oada technical reference: Building interoperable agricultural data sys- tems,” OpenAg Foundation, Technical Report, 2023

work page 2023

[4] [4]

Improving decision support systems with machine learning: Identifying barriers to adoption,

S. Brugler, M. Gardezi, A. Dadkhah, D. M. Rizzo, A. Zia, and S. A. Clay, “Improving decision support systems with machine learning: Identifying barriers to adoption,” Agronomy Journal, vol. 116, no. 3, pp. 1229–1236, 2024

work page 2024

[5] [5]

Deep learning for image-based cassava disease detection,

A. Ramcharan, K. Baranowski, P. McCloskey, B. Ahmed, J. Legg, and D. P. Hughes, “Deep learning for image-based cassava disease detection,” Frontiers in plant science , vol. 8, p. 1852, 2017

work page 2017

[6] [6]

Transfer learning for crop type classification: The agricultural breakthrough with ai,

H. F ARHAN, L. Di, and C. Zhang, “Transfer learning for crop type classification: The agricultural breakthrough with ai,” Journal of Student-Scientists’ Research , vol. 5, 2023

work page 2023

[7] [7]

A comprehensive review of the data and knowledge graphs approaches in bioinfor- matics,

Y. Galluzzo, “A comprehensive review of the data and knowledge graphs approaches in bioinfor- matics,” Computer Science and Information Systems , no. 00, pp. 27–27, 2024

work page 2024

[8] [8]

Knowledge graphs: Opportunities and chal- lenges,

C. Peng, F. Xia, M. Naseriparsa, and F. Osborne, “Knowledge graphs: Opportunities and chal- lenges,” Artificial Intelligence Review , vol. 56, no. 11, pp. 13 071–13 102, 2023

work page 2023