Publications

@article{dinkins_localized_2026,

title = {Localized Near Surface Temperature Inversion Forecasting Using Long Short-Term Memory},

author = {Taylor Dinkins and Weng-Keen Wong and Basavaraj Amogi and Paola Pesantez-Cabrera and Jaitun Patel and Lav Khot and Alan Fern},

url = {https://ojs.aaai.org/index.php/AAAI/article/view/41462},

doi = {10.1609/aaai.v40i47.41462},

issn = {2374-3468},

year  = {2026},

date = {2026-03-01},

urldate = {2026-03-01},

journal = {Proceedings of the AAAI Conference on Artificial Intelligence},

volume = {40},

number = {47},

pages = {40249\textendash40257},

abstract = {Near surface temperature inversions are periods in which a low layer of warm air is trapped between cooler air higher up in the atmosphere and dense cooler air below it near the surface level. By causing cooler air to pool near the surface level, inversions can have detrimental effects for crop growers, including frost, increased moisture, and pesticide drift. As a result, predicting the occurrence and magnitude of these inversions yields substantial benefits for growers. We introduce a Long Short-Term Memory (LSTM) model for temperature inversion forecasting that is able to effectively predict localized, near surface temperature inversions in advance such that growers can take actions to mitigate the detrimental effects. We show a substantial performance gain over a deployed temperature inversion forecasting system, and include a series of ablations that show the benefit of using publicly available terrain-specific feature information when modeling inversions at this scale.},

keywords = {AI, Farm Ops},

pubstate = {published},

tppubtype = {article}

}

@article{goebel_budgeted_2026,

title = {Budgeted Online Active Learning with Expert Advice and Episodic Priors},

author = {Kristen Goebel and William Solow and Paola Pesantez-Cabrera and Markus Keller and Alan Fern},

url = {https://ojs.aaai.org/index.php/AAAI/article/view/41191},

doi = {10.1609/aaai.v40i45.41191},

issn = {2374-3468},

year  = {2026},

date = {2026-03-01},

urldate = {2026-03-01},

journal = {Proceedings of the AAAI Conference on Artificial Intelligence},

volume = {40},

number = {45},

pages = {38496\textendash38504},

abstract = {This paper introduces a novel approach to budgeted online active learning from finite-horizon data streams with extremely limited labeling budgets. In agricultural applications, such streams might include daily weather data over a growing season, and labels require costly measurements of weather-dependent plant characteristics. Our method integrates two key sources of prior information: a collection of preexisting expert predictors and episodic behavioral knowledge of the experts based on unlabeled data streams. Unlike previous research on online active learning with experts, our work simultaneously considers query budgets, finite horizons, and episodic knowledge, enabling effective learning in applications with severely limited labeling capacity. We demonstrate the utility of our approach through experiments on various prediction problems derived from both a realistic agricultural crop simulator and real-world data from multiple grape cultivars. The results show that our method significantly outperforms baseline expert predictions, uniform query selection, and existing approaches that consider budgets and limited horizons but neglect episodic knowledge, even under highly constrained labeling budgets.},

keywords = {AI, Farm Ops},

pubstate = {published},

tppubtype = {article}

}

@article{roumeliotis_agentic_2026b,

title = {Agentic AI With Orchestrator-Agent Trust: A Modular Visual Classification Framework With Trust-Aware Orchestration and RAG-Based Reasoning},

author = {Konstantinos I. Roumeliotis and Ranjan Sapkota and Manoj Karkee and Nikolaos D. Tselikas},

url = {https://ieeexplore.ieee.org/document/11373381/},

doi = {10.1109/ACCESS.2026.3662282},

issn = {2169-3536},

year  = {2026},

date = {2026-01-01},

urldate = {2026-01-01},

journal = {IEEE Access},

volume = {14},

pages = {26965\textendash26982},

abstract = {Modern Artificial Intelligence (AI) increasingly relies on multi-agent architectures that blend visual and language understanding. Yet, a pressing challenge remains: How can we trust these agents especially in zero-shot settings with no fine-tuning? We introduce a novel modular Agentic AI visual classification framework that integrates generalist multimodal agents with a non-visual reasoning orchestrator and a Retrieval-Augmented Generation (RAG) module. Applied to apple leaf disease diagnosis, we benchmark three configurations: (I) zero-shot with confidence-based orchestration, (II) fine-tuned agents with improved performance, and (III) trust-calibrated orchestration enhanced by CLIP-based image retrieval and re-evaluation loops. Using confidence calibration metrics (ECE, OCR, CCC), the orchestrator modulates trust across agents. Our results demonstrate a 77.94% accuracy improvement in the zero-shot setting using trust-aware orchestration and RAG, achieving 85.63% overall. GPT-4o showed better calibration, while Qwen-2.5-VL displayed overconfidence. Furthermore, image-RAG grounded predictions with visually similar cases, enabling correction of agent overconfidence via iterative re-evaluation. The proposed system separates perception (vision agents) from meta-reasoning (orchestrator), enabling scalable and interpretable multi-agent AI. This blueprint illustrates how Agentic AI can deliver trustworthy, modular, and transparent reasoning, and is extensible to diagnostics, biology, and other trust-critical domains. In doing so, we highlight Agentic AI not just as an architecture but as a paradigm for building reliable multi-agent intelligence. All models, prompts, results, and system components including the complete software source code are openly released to support reproducibility, transparency, and community benchmarking at our Github page.},

keywords = {AI, Labor},

pubstate = {published},

tppubtype = {article}

}

@article{hamid_inclusive_2026,

title = {Inclusive Design of AI’s Explanations: Just for Those Previously Left Out?},

author = {Md Montaser Hamid and Fatima A. Moussaoui and Jimena Noa Guevara and Andrew Anderson and Puja Agarwal and Jonathan Dodge and Margaret Burnett},

url = {https://dl.acm.org/doi/10.1145/3772074},

doi = {10.1145/3772074},

issn = {2160-6455},

year  = {2026},

date = {2026-01-01},

urldate = {2026-01-01},

journal = {ACM Transactions on Interactive Intelligent Systems},

volume = {16},

number = {1},

pages = {8:1\textendash8:44},

abstract = {AbstractMotivations. Explainable AI (XAI) systems aim to improve users’ understanding of AI, but XAI research has shown that many XAI explanations serve some users well while failing others. In non-AI systems, software practitioners have used inclusive design approaches to address similar problems, sometimes creating “curb-cut” improvements that benefit both underserved users and everyone else. This raises the possibility that inclusive design approaches can bring similar curb-cut improvements to AI explanations. Objectives. Our objective was to investigate possible curb-cut effects of inclusivity-driven fixes an AI product team made using an inclusive design approach (GenderMag) to improve their XAI prototype. Methods. We ran a between-subject study with 69 participants who had no formal AI background. 34 participants used the original version of the XAI prototype and the rest used the version with the AI team’s inclusivity fixes. We then compared the two groups’ mental model concepts scores and prediction accuracy, and the two prototypes’ inclusivity. Results. Our investigation produced four main results. First, the AI team’s inclusivity fixes were overall effective, resulting in overall better conceptual mental models with the new prototype. Further (second), the AI team’s inclusivity fixes were particularly beneficial to the underserved population’s conceptual mental models\textemdashwhich, together with the first result, constitutes a curb-cut effect. However (third), the inclusivity fixes did not improve participants’ prediction accuracy scores. Instead, it appears to have harmed them overall\textemdasha “curb-fence” effect (opposite of a curb-cut effect). Finally (fourth), the AI team’s fixes improved equity, reducing the gender gap by 45%.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{raza_trism_2026,

title = {TRiSM for Agentic AI: A review of Trust, Risk, and Security Management in LLM-based Agentic Multi-Agent Systems},

author = {Shaina Raza and Ranjan Sapkota and Manoj Karkee and Christos Emmanouilidis},

url = {https://www.sciencedirect.com/science/article/pii/S2666651026000069},

doi = {https://doi.org/10.1016/j.aiopen.2026.02.006},

issn = {2666-6510},

year  = {2026},

date = {2026-01-01},

urldate = {2026-01-01},

journal = {AI Open},

volume = {7},

pages = {71\textendash95},

abstract = {Agentic AI systems, built upon large language models (LLMs) and deployed in multi-agent configurations, are redefining intelligence, autonomy, collaboration, and decision-making across enterprise and societal domains. This review presents a structured analysis of Trust, Risk, and Security Management (TRiSM) in the context of LLM-based Agentic Multi-Agent Systems (AMAS). We begin by examining the conceptual foundations of Agentic AI and highlight its architectural distinctions from traditional AI agents. We then adapt and extend the AI TRiSM framework for Agentic AI, structured around key pillars: Explainability, ModelOps, Security, Privacy and their Lifecycle Governance, each contextualized to the challenges of AMAS. A risk taxonomy is proposed to capture the unique threats and vulnerabilities of Agentic AI, ranging from coordination failures to prompt-based adversarial manipulation. To make coordination and tool use measurable in practice, we propose two metrics: the Component Synergy Score (CSS), which captures inter-agent enablement, and the Tool Utilization Efficacy (TUE), which evaluates whether tools are invoked correctly and efficiently. We further discuss strategies for improving explainability in Agentic AI, as well as approaches to enhancing security and privacy through encryption, adversarial robustness, and regulatory compliance. The review concludes with a research roadmap for the responsible development and deployment of Agentic AI, highlighting key directions to align emerging systems with TRiSM principles-ensuring safety, transparency, and accountability in their operation.},

keywords = {Adversarial robustness, Agentic AI, AI agents, AI governance, AI safety, Application security, Explainability, Human-in-the-Loop, LLM-based multi-agent systems, Model Privacy, ModelOps, Privacy-preserving AI, Risk management, TRiSM, Trustworthy AI},

pubstate = {published},

tppubtype = {article}

}

@article{sapkota_comprehensive_2026,

title = {Comprehensive performance evaluation of YOLOv12, YOLO11, YOLOv10, YOLOv9 and YOLOv8 on detecting and counting fruitlet in complex orchard environments},

author = {Ranjan Sapkota and Zhichao Meng and Martin Churuvija and Xiaoqiang Du and Zenghong Ma and Manoj Karkee},

url = {https://www.sciencedirect.com/science/article/pii/S2949798126000050},

doi = {https://doi.org/10.1016/j.agrcom.2026.100125},

issn = {2949-7981},

year  = {2026},

date = {2026-01-01},

urldate = {2026-01-01},

journal = {Agriculture Communications},

volume = {4},

number = {1},

pages = {100125},

abstract = {This study systematically conducted an extensive real-world evaluation of all configurations of You Only Look Once (YOLO)-based object detection algorithms, including YOLOv8, YOLOv9, YOLOv10, YOLO11, and YOLOv12. Models were assessed using precision, recall, mean Average Precision at 50 % Intersection over Union (mAP@50), and computational efficiency across pre-processing, inference, and post-processing stages for detecting immature green fruitlets in commercial orchards. Field-level fruitlet counting was also validated using images captured with both Intel RealSense and iPhone 14 Pro Max sensors. YOLOv12l achieved the highest recall (0.900), while YOLOv10x and YOLOv9 GELAN-c reported the top precision scores of 0.908 and 0.903, respectively. YOLOv9 GELAN-base and GELAN-e achieved the highest mAP@50 (0.935), followed by YOLO11s (0.933) and YOLOv12l (0.931). In counting validation, YOLO11n demonstrated superior accuracy, with RMSE values of 4.51\textendash4.96 and MAE values of 3.85\textendash7.73 across four apple varieties. Sensor-specific training on Intel RealSense further improved detection performance. YOLO11n also recorded the fastest inference speed (2.4 ms), outperforming YOLOv8n, YOLOv9 GELAN-s, YOLOv10n, and YOLOv12n, affirming its suitability for real-time orchard applications.},

keywords = {agricultural automation, Fruitlet detection, Object detection, YOLO comparison, You Only Look Once (YOLO)},

pubstate = {published},

tppubtype = {article}

}

@article{roumeliotis_agentic_2026,

title = {Agentic AI With Orchestrator-Agent Trust: A Modular Visual Classification Framework With Trust-Aware Orchestration and RAG-Based Reasoning},

author = {Konstantinos I. Roumeliotis and Ranjan Sapkota and Manoj Karkee and Nikolaos D. Tselikas},

url = {https://ieeexplore.ieee.org/document/11373381/},

doi = {10.1109/ACCESS.2026.3662282},

issn = {2169-3536},

year  = {2026},

date = {2026-01-01},

urldate = {2026-01-01},

journal = {IEEE Access},

volume = {14},

pages = {26965\textendash26982},

abstract = {Modern Artificial Intelligence (AI) increasingly relies on multi-agent architectures that blend visual and language understanding. Yet, a pressing challenge remains: How can we trust these agents especially in zero-shot settings with no fine-tuning? We introduce a novel modular Agentic AI visual classification framework that integrates generalist multimodal agents with a non-visual reasoning orchestrator and a Retrieval-Augmented Generation (RAG) module. Applied to apple leaf disease diagnosis, we benchmark three configurations: (I) zero-shot with confidence-based orchestration, (II) fine-tuned agents with improved performance, and (III) trust-calibrated orchestration enhanced by CLIP-based image retrieval and re-evaluation loops. Using confidence calibration metrics (ECE, OCR, CCC), the orchestrator modulates trust across agents. Our results demonstrate a 77.94% accuracy improvement in the zero-shot setting using trust-aware orchestration and RAG, achieving 85.63% overall. GPT-4o showed better calibration, while Qwen-2.5-VL displayed overconfidence. Furthermore, image-RAG grounded predictions with visually similar cases, enabling correction of agent overconfidence via iterative re-evaluation. The proposed system separates perception (vision agents) from meta-reasoning (orchestrator), enabling scalable and interpretable multi-agent AI. This blueprint illustrates how Agentic AI can deliver trustworthy, modular, and transparent reasoning, and is extensible to diagnostics, biology, and other trust-critical domains. In doing so, we highlight Agentic AI not just as an architecture but as a paradigm for building reliable multi-agent intelligence. All models, prompts, results, and system components including the complete software source code are openly released to support reproducibility, transparency, and community benchmarking at our Github page.},

keywords = {Accuracy, Adaptation models, Agentic AI, Artificial intelligence, Calibration, Cognition, Costs, orchestrator agent trust, Retrieval augmented generation, retrieval augmented reasoning, Training, trust orchestration, visual classification, Visualization},

pubstate = {published},

tppubtype = {article}

}

@article{cravetz_slip_2025b,

title = {Slip detection for compliant robotic hands using inertial signals and deep learning},

author = {Miranda Cravetz and Purva Vyas and Cindy Grimm and Joseph R. Davidson},

url = {https://www.frontiersin.org/journals/robotics-and-ai/articles/10.3389/frobt.2025.1698591/full},

doi = {10.3389/frobt.2025.1698591},

issn = {2296-9144},

year  = {2025},

date = {2025-12-01},

urldate = {2025-12-01},

journal = {Frontiers in Robotics and AI},

volume = {12},

publisher = {Frontiers},

abstract = {When a passively compliant hand grasps an object, slip events are often accompanied by flexion or extension of the finger or finger joints. This paper investigates whether a combination of orientation change and slip-induced vibration at the fingertip, as sensed by an inertial measurement unit (IMU), can be used as a slip indicator. Using a tendon-driven hand, which achieves passive compliance through underactuation, we performed 195 manipulation trials involving both slip and non-slip conditions. We then labeled this data automatically using motion-tracking data, and trained a convolutional neural network (CNN) to detect the slip events. Our results show that slip can be successfully detected from IMU data, even in the presence of other disturbances. This remains the case when deploying the trained network on data from a different gripper performing a new manipulation task on a previously unseen object.},

keywords = {AI, Labor},

pubstate = {published},

tppubtype = {article}

}