2025
1.
Kristen Goebel; Paola Pesantez-Cabrera; Markus Keller; Alan Fern
Transfer Learning via Auxiliary Labels with Application to Cold-Hardiness Prediction Working paper
2025, (arXiv:2504.13142 [cs]).
Abstract | Links | BibTeX | Tags:
@workingpaper{goebel_transfer_2025,
title = {Transfer Learning via Auxiliary Labels with Application to Cold-Hardiness Prediction},
author = {Kristen Goebel and Paola Pesantez-Cabrera and Markus Keller and Alan Fern},
url = {http://arxiv.org/abs/2504.13142},
doi = {10.48550/arXiv.2504.13142},
year = {2025},
date = {2025-04-01},
urldate = {2025-04-01},
publisher = {arXiv},
abstract = {Cold temperatures can cause significant frost damage to fruit crops depending on their resilience, or cold hardiness, which changes throughout the dormancy season. This has led to the development of predictive cold-hardiness models, which help farmers decide when to deploy expensive frost-mitigation measures. Unfortunately, cold-hardiness data for model training is only available for some fruit cultivars due to the need for specialized equipment and expertise. Rather, farmers often do have years of phenological data (e.g. date of budbreak) that they regularly collect for their crops. In this work, we introduce a new transfer-learning framework, Transfer via Auxiliary Labels (TAL), that allows farmers to leverage the phenological data to produce more accurate cold-hardiness predictions, even when no cold-hardiness data is available for their specific crop. The framework assumes a set of source tasks (cultivars) where each has associated primary labels (cold hardiness) and auxiliary labels (phenology). However, the target task (new cultivar) is assumed to only have the auxiliary labels. The goal of TAL is to predict primary labels for the target task via transfer from the source tasks. Surprisingly, despite the vast literature on transfer learning, to our knowledge, the TAL formulation has not been previously addressed. Thus, we propose several new TAL approaches based on model selection and averaging that can leverage recent deep multi-task models for cold-hardiness prediction. Our results on real-world cold-hardiness and phenological data for multiple grape cultivars demonstrate that TAL can leverage the phenological data to improve cold-hardiness predictions in the absence of cold-hardiness data.},
note = {arXiv:2504.13142 [cs]},
keywords = {},
pubstate = {published},
tppubtype = {workingpaper}
}
Cold temperatures can cause significant frost damage to fruit crops depending on their resilience, or cold hardiness, which changes throughout the dormancy season. This has led to the development of predictive cold-hardiness models, which help farmers decide when to deploy expensive frost-mitigation measures. Unfortunately, cold-hardiness data for model training is only available for some fruit cultivars due to the need for specialized equipment and expertise. Rather, farmers often do have years of phenological data (e.g. date of budbreak) that they regularly collect for their crops. In this work, we introduce a new transfer-learning framework, Transfer via Auxiliary Labels (TAL), that allows farmers to leverage the phenological data to produce more accurate cold-hardiness predictions, even when no cold-hardiness data is available for their specific crop. The framework assumes a set of source tasks (cultivars) where each has associated primary labels (cold hardiness) and auxiliary labels (phenology). However, the target task (new cultivar) is assumed to only have the auxiliary labels. The goal of TAL is to predict primary labels for the target task via transfer from the source tasks. Surprisingly, despite the vast literature on transfer learning, to our knowledge, the TAL formulation has not been previously addressed. Thus, we propose several new TAL approaches based on model selection and averaging that can leverage recent deep multi-task models for cold-hardiness prediction. Our results on real-world cold-hardiness and phenological data for multiple grape cultivars demonstrate that TAL can leverage the phenological data to improve cold-hardiness predictions in the absence of cold-hardiness data.
2.
Ozmen Erkin Kokten; Raviv Raich
Maximum Likelihood Estimation of Stable ARX Models using Randomized Coordinate Descent Proceedings Article
In: ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 1–5, 2025, (ISSN: 2379-190X).
Abstract | Links | BibTeX | Tags: autoregressive model with exogenous variables, Convergence, coordinate descent, Data models, Finance, Numerical models, Numerical stability, optimization, parameter estimation, Signal processing, Speech processing, stability, Stability criteria
@inproceedings{kokten_maximum_2025,
title = {Maximum Likelihood Estimation of Stable ARX Models using Randomized Coordinate Descent},
author = {Ozmen Erkin Kokten and Raviv Raich},
url = {https://ieeexplore.ieee.org/abstract/document/10888613/authors},
doi = {10.1109/ICASSP49660.2025.10888613},
year = {2025},
date = {2025-04-01},
urldate = {2025-04-01},
booktitle = {ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
pages = {1\textendash5},
abstract = {Autoregressive models play an important role in a variety of applications including finance, engineering, sciences, and agriculture. While for some models (e.g., physics-based models) parameters are known, in other domains the parameters may not be available. This paper deals with the estimation of the parameters of an autoregressive model with exogenous variables. A significant body of literature has explored autoregressive model estimation across different estimation criteria, data availability, and parameterization; however, limited attention has been given to the estimation problem under stability constraints. The incorporation of stability constraints often results in increased computational complexity. As an efficient alternative, we propose to estimate stable ARX parameters using randomized coordinate descent. To demonstrate the efficiency of the proposed approach, we present an empirical convergence study and compare our approach to a state-of-the-art alternative.},
note = {ISSN: 2379-190X},
keywords = {autoregressive model with exogenous variables, Convergence, coordinate descent, Data models, Finance, Numerical models, Numerical stability, optimization, parameter estimation, Signal processing, Speech processing, stability, Stability criteria},
pubstate = {published},
tppubtype = {inproceedings}
}
Autoregressive models play an important role in a variety of applications including finance, engineering, sciences, and agriculture. While for some models (e.g., physics-based models) parameters are known, in other domains the parameters may not be available. This paper deals with the estimation of the parameters of an autoregressive model with exogenous variables. A significant body of literature has explored autoregressive model estimation across different estimation criteria, data availability, and parameterization; however, limited attention has been given to the estimation problem under stability constraints. The incorporation of stability constraints often results in increased computational complexity. As an efficient alternative, we propose to estimate stable ARX parameters using randomized coordinate descent. To demonstrate the efficiency of the proposed approach, we present an empirical convergence study and compare our approach to a state-of-the-art alternative.
3.
Mohsen Farajijalal; Ali Abedi; Cristian Manzo; Amir Kouravand; Mohammadmehdi Maharlooei; Arash Toudeshki; Reza Ehsani
Assessing Crucial Shaking Parameters in the Mechanical Harvesting of Nut Trees: A Review Journal Article
In: Horticulturae, vol. 11, no. 4, pp. 392, 2025, ISSN: 2311-7524, (Number: 4 Publisher: Multidisciplinary Digital Publishing Institute).
Abstract | Links | BibTeX | Tags: amplitude, frequency, harvest efficiency, shaker, vibration
@article{farajijalal_assessing_2025,
title = {Assessing Crucial Shaking Parameters in the Mechanical Harvesting of Nut Trees: A Review},
author = {Mohsen Farajijalal and Ali Abedi and Cristian Manzo and Amir Kouravand and Mohammadmehdi Maharlooei and Arash Toudeshki and Reza Ehsani},
url = {https://www.mdpi.com/2311-7524/11/4/392},
doi = {10.3390/horticulturae11040392},
issn = {2311-7524},
year = {2025},
date = {2025-04-01},
urldate = {2025-04-01},
journal = {Horticulturae},
volume = {11},
number = {4},
pages = {392},
abstract = {Finding appropriate shaking parameters is crucial in designing effective mechanical harvesters. The maximum fruit removal can be achieved when the machine operator properly adjusts the amplitude and frequency for shaking each tree. This review covers the progress in research and development over the past decades on using mechanical harvesters for nut trees, such as almonds, pistachios, walnuts, and hickories, with a specific focus on the natural frequency of individual trees. Furthermore, the reported values of shaking frequency and amplitude from previous studies were discussed and compared, along with frequency calculation approaches based on various shaking mechanisms. Additionally, other parameters, such as clamping force, height, and shaking amplitude, were investigated to determine optimal values for minimizing tree damage. This review emphasizes that the tree’s diameter, height, and canopy morphology should be the primary factors considered when estimating the optimal shaking frequency for nut trees. It also highlights that, to date, the shaking amplitude, frequency, and duration set by field managers or machine operators tend to remain consistent for all trees, which can limit harvesting efficiency. The findings suggest that selecting these parameters uniformly across all trees may not result in efficient fruit removal for individual trees. However, with the assistance of modern computing technology and its adaptation for in-field applications, it is feasible to determine the optimal shaking frequency for each tree mathematically. This approach can maximize fruit removal rates while minimizing tree damage. Finally, the review suggests that improving existing harvesting machines by incorporating better vibratory patterns could offer benefits such as enhanced productivity, reduced labor costs, and decreased permanent tree damage.},
note = {Number: 4
Publisher: Multidisciplinary Digital Publishing Institute},
keywords = {amplitude, frequency, harvest efficiency, shaker, vibration},
pubstate = {published},
tppubtype = {article}
}
Finding appropriate shaking parameters is crucial in designing effective mechanical harvesters. The maximum fruit removal can be achieved when the machine operator properly adjusts the amplitude and frequency for shaking each tree. This review covers the progress in research and development over the past decades on using mechanical harvesters for nut trees, such as almonds, pistachios, walnuts, and hickories, with a specific focus on the natural frequency of individual trees. Furthermore, the reported values of shaking frequency and amplitude from previous studies were discussed and compared, along with frequency calculation approaches based on various shaking mechanisms. Additionally, other parameters, such as clamping force, height, and shaking amplitude, were investigated to determine optimal values for minimizing tree damage. This review emphasizes that the tree’s diameter, height, and canopy morphology should be the primary factors considered when estimating the optimal shaking frequency for nut trees. It also highlights that, to date, the shaking amplitude, frequency, and duration set by field managers or machine operators tend to remain consistent for all trees, which can limit harvesting efficiency. The findings suggest that selecting these parameters uniformly across all trees may not result in efficient fruit removal for individual trees. However, with the assistance of modern computing technology and its adaptation for in-field applications, it is feasible to determine the optimal shaking frequency for each tree mathematically. This approach can maximize fruit removal rates while minimizing tree damage. Finally, the review suggests that improving existing harvesting machines by incorporating better vibratory patterns could offer benefits such as enhanced productivity, reduced labor costs, and decreased permanent tree damage.
4.
Adam Weingram; Carolyn Cui; Stephanie Lin; Samuel Munoz; Toby Jacob; Joshua Viers; Xiaoyi Lu
A definition and taxonomy of digital twins: case studies with machine learning and scientific applications Journal Article
In: Frontiers in High Performance Computing, vol. 3, 2025, ISSN: 2813-7337, (Publisher: Frontiers).
@article{weingram_definition_2025,
title = {A definition and taxonomy of digital twins: case studies with machine learning and scientific applications},
author = {Adam Weingram and Carolyn Cui and Stephanie Lin and Samuel Munoz and Toby Jacob and Joshua Viers and Xiaoyi Lu},
url = {https://www.frontiersin.org/journals/high-performance-computing/articles/10.3389/fhpcp.2025.1536501/full},
doi = {10.3389/fhpcp.2025.1536501},
issn = {2813-7337},
year = {2025},
date = {2025-03-01},
urldate = {2025-03-01},
journal = {Frontiers in High Performance Computing},
volume = {3},
note = {Publisher: Frontiers},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
5.
William Solow; Sandhya Saisubramanian; Alan Fern
WOFOSTGym: A Crop Simulator for Learning Annual and Perennial Crop Management Strategies Miscellaneous
2025, (arXiv:2502.19308 [cs]).
Abstract | Links | BibTeX | Tags:
@misc{solow_wofostgym_2025,
title = {WOFOSTGym: A Crop Simulator for Learning Annual and Perennial Crop Management Strategies},
author = {William Solow and Sandhya Saisubramanian and Alan Fern},
url = {http://arxiv.org/abs/2502.19308},
doi = {10.48550/arXiv.2502.19308},
year = {2025},
date = {2025-02-01},
urldate = {2025-02-01},
publisher = {arXiv},
abstract = {We introduce WOFOSTGym, a novel crop simulation environment designed to train reinforcement learning (RL) agents to optimize agromanagement decisions for annual and perennial crops in single and multi-farm settings. Effective crop management requires optimizing yield and economic returns while minimizing environmental impact, a complex sequential decision-making problem well suited for RL. However, the lack of simulators for perennial crops in multi-farm contexts has hindered RL applications in this domain. Existing crop simulators also do not support multiple annual crops. WOFOSTGym addresses these gaps by supporting 23 annual crops and two perennial crops, enabling RL agents to learn diverse agromanagement strategies in multi-year, multi-crop, and multi-farm settings. Our simulator offers a suite of challenging tasks for learning under partial observability, non-Markovian dynamics, and delayed feedback. WOFOSTGym's standard RL interface allows researchers without agricultural expertise to explore a wide range of agromanagement problems. Our experiments demonstrate the learned behaviors across various crop varieties and soil types, highlighting WOFOSTGym's potential for advancing RL-driven decision support in agriculture.},
note = {arXiv:2502.19308 [cs]},
keywords = {},
pubstate = {published},
tppubtype = {misc}
}
We introduce WOFOSTGym, a novel crop simulation environment designed to train reinforcement learning (RL) agents to optimize agromanagement decisions for annual and perennial crops in single and multi-farm settings. Effective crop management requires optimizing yield and economic returns while minimizing environmental impact, a complex sequential decision-making problem well suited for RL. However, the lack of simulators for perennial crops in multi-farm contexts has hindered RL applications in this domain. Existing crop simulators also do not support multiple annual crops. WOFOSTGym addresses these gaps by supporting 23 annual crops and two perennial crops, enabling RL agents to learn diverse agromanagement strategies in multi-year, multi-crop, and multi-farm settings. Our simulator offers a suite of challenging tasks for learning under partial observability, non-Markovian dynamics, and delayed feedback. WOFOSTGym's standard RL interface allows researchers without agricultural expertise to explore a wide range of agromanagement problems. Our experiments demonstrate the learned behaviors across various crop varieties and soil types, highlighting WOFOSTGym's potential for advancing RL-driven decision support in agriculture.
6.
Dawood Ahmed; Ranjan Sapkota; Martin Churuvija; Manoj Karkee
Estimating optimal crop-load for individual branches in apple tree canopies using YOLOv8 Journal Article
In: Computers and Electronics in Agriculture, vol. 229, pp. 109697, 2025, ISSN: 0168-1699.
Abstract | Links | BibTeX | Tags:
@article{ahmed_estimating_2025,
title = {Estimating optimal crop-load for individual branches in apple tree canopies using YOLOv8},
author = {Dawood Ahmed and Ranjan Sapkota and Martin Churuvija and Manoj Karkee},
url = {https://www.sciencedirect.com/science/article/pii/S0168169924010883},
doi = {10.1016/j.compag.2024.109697},
issn = {0168-1699},
year = {2025},
date = {2025-02-01},
urldate = {2025-02-01},
journal = {Computers and Electronics in Agriculture},
volume = {229},
pages = {109697},
abstract = {Shortage of labor in fruit crop production has become a significant challenge in recent years. Therefore, mechanized and automated machines have emerged as promising alternatives to labor-intensive orchard operations such as harvesting, pruning, and thinning. One of the key aspects of the automated machines in accomplishing these tasks is their ability to identify tree canopy parts such as trunk and branches and estimate their geometric and topological parameters such as branch diameter, branch length, branch angles, and spacing between branches. By utilizing geometric parameters such as branch diameter, length, and orientation, researchers then can develop automated pruning and thinning systems that make more effective decisions to achieve optimal fruit yield and quality by accurately estimating the desired crop-load. In this study, we propose a machine vision system for estimating one of the canopy parameters in apple orchards: branch diameter. This parameter was used to estimate the optimal number of fruit that individual branches could bear in a commercial orchard, which provides a basis for robotic pruning, flower thinning, and fruitlet thinning so that desired fruit yield and quality could be achieved. Utilizing color and depth information collected with an RGB-D sensor (Azure Kinect DK, Microsoft, Redmond, WA), a YOLOv8-based instance segmentation technique was developed to identify trunks and branches of apple trees in the dormant season. We then applied a Principal Component Analysis (PCA) technique to estimate branch orientation, which was subsequently utilized to estimate branch diameter. The estimated branch diameter was used to calculate the Limb Cross Sectional Area (LCSA), which was then used to estimate optimal crop-load, as a larger LCSA indicates a higher potential fruit-bearing capacity of the branch. With this approach, Root Mean Squared Error (RMSE) for branch diameter estimation was calculated to be 2.06 mm (relative RMSE 10.82%) and the same for crop-load estimation (Number of fruits per branch) to be 3.93 (relative RMSE 22.25%). Our study demonstrated a promising workflow with a high level of performance in identifying and sizing branches of apple trees in a dynamic orchard environment and integrating farm management practices into automated decision-making for optimizing crop-load in apple orchards.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Shortage of labor in fruit crop production has become a significant challenge in recent years. Therefore, mechanized and automated machines have emerged as promising alternatives to labor-intensive orchard operations such as harvesting, pruning, and thinning. One of the key aspects of the automated machines in accomplishing these tasks is their ability to identify tree canopy parts such as trunk and branches and estimate their geometric and topological parameters such as branch diameter, branch length, branch angles, and spacing between branches. By utilizing geometric parameters such as branch diameter, length, and orientation, researchers then can develop automated pruning and thinning systems that make more effective decisions to achieve optimal fruit yield and quality by accurately estimating the desired crop-load. In this study, we propose a machine vision system for estimating one of the canopy parameters in apple orchards: branch diameter. This parameter was used to estimate the optimal number of fruit that individual branches could bear in a commercial orchard, which provides a basis for robotic pruning, flower thinning, and fruitlet thinning so that desired fruit yield and quality could be achieved. Utilizing color and depth information collected with an RGB-D sensor (Azure Kinect DK, Microsoft, Redmond, WA), a YOLOv8-based instance segmentation technique was developed to identify trunks and branches of apple trees in the dormant season. We then applied a Principal Component Analysis (PCA) technique to estimate branch orientation, which was subsequently utilized to estimate branch diameter. The estimated branch diameter was used to calculate the Limb Cross Sectional Area (LCSA), which was then used to estimate optimal crop-load, as a larger LCSA indicates a higher potential fruit-bearing capacity of the branch. With this approach, Root Mean Squared Error (RMSE) for branch diameter estimation was calculated to be 2.06 mm (relative RMSE 10.82%) and the same for crop-load estimation (Number of fruits per branch) to be 3.93 (relative RMSE 22.25%). Our study demonstrated a promising workflow with a high level of performance in identifying and sizing branches of apple trees in a dynamic orchard environment and integrating farm management practices into automated decision-making for optimizing crop-load in apple orchards.
2024
7.
Ozmen Erkin Kokten; Raviv Raich; James Holmes; Alan Fern
Learning Extended Forecasts of Soil Water Content via Physically-Inspired Autoregressive Models Proceedings Article
In: 2024 International Conference on Machine Learning and Applications (ICMLA), pp. 400–407, 2024, (ISSN: 1946-0759).
Abstract | Links | BibTeX | Tags: autoregressive training, non-linear state-space models, Pipelines, Predictive models, Soil measurements, State-space methods, Stress, teacher-forcing, time-series, Training, Training data, Weather forecasting
@inproceedings{kokten_learning_2024,
title = {Learning Extended Forecasts of Soil Water Content via Physically-Inspired Autoregressive Models},
author = {Ozmen Erkin Kokten and Raviv Raich and James Holmes and Alan Fern},
url = {https://ieeexplore.ieee.org/abstract/document/10903312},
doi = {10.1109/ICMLA61862.2024.00060},
year = {2024},
date = {2024-12-01},
urldate = {2024-12-01},
booktitle = {2024 International Conference on Machine Learning and Applications (ICMLA)},
pages = {400\textendash407},
abstract = {Vine stress resulting from soil water content (SWC) restrictions allows growers to improve grape and subsequent wine quality. In this work, we consider learning models that can forecast SWC to assist growers' irrigation decisions. In particular, we investigate training auto-regressive recurrent neural networks to make multi-day hourly forecasts of SWC based on historical data from soil-moisture sensors, irrigation sched-ules, and evapotranspiration estimates. Our work addresses two practical challenges in training such models. First, trained auto-regressive models are prone to error propagation, which quickly degrades longer-term forecasts. Second, it is difficult to learn the underlying causal relationship between irrigation and soil moisture due to the training data having limited coverage of the primary control input, irrigation. We propose a training strategy that combines one-step teacher forcing loss with a loss over multi-step autoregressive predictions and novel regularization terms to ensure SWC forecasts align with scientific models, effectively addressing the key challenges. We present results from five irrigation blocks with two cultivars, using datasets ranging from 2947 to 4784 hourly measurements of SWC, irrigation, and weather. Our methodology achieves precise SWC predictions and generates realistic forecasts for untrained irrigation scenarios.},
note = {ISSN: 1946-0759},
keywords = {autoregressive training, non-linear state-space models, Pipelines, Predictive models, Soil measurements, State-space methods, Stress, teacher-forcing, time-series, Training, Training data, Weather forecasting},
pubstate = {published},
tppubtype = {inproceedings}
}
Vine stress resulting from soil water content (SWC) restrictions allows growers to improve grape and subsequent wine quality. In this work, we consider learning models that can forecast SWC to assist growers' irrigation decisions. In particular, we investigate training auto-regressive recurrent neural networks to make multi-day hourly forecasts of SWC based on historical data from soil-moisture sensors, irrigation sched-ules, and evapotranspiration estimates. Our work addresses two practical challenges in training such models. First, trained auto-regressive models are prone to error propagation, which quickly degrades longer-term forecasts. Second, it is difficult to learn the underlying causal relationship between irrigation and soil moisture due to the training data having limited coverage of the primary control input, irrigation. We propose a training strategy that combines one-step teacher forcing loss with a loss over multi-step autoregressive predictions and novel regularization terms to ensure SWC forecasts align with scientific models, effectively addressing the key challenges. We present results from five irrigation blocks with two cultivars, using datasets ranging from 2947 to 4784 hourly measurements of SWC, irrigation, and weather. Our methodology achieves precise SWC predictions and generates realistic forecasts for untrained irrigation scenarios.
8.
Yassine Chemingui; Aryan Deshwal; Honghao Wei; Alan Fern; Janardhan Rao Doppa
Constraint-Adaptive Policy Switching for Offline Safe Reinforcement Learning Proceedings Article
In: arXiv, 2024, (arXiv:2412.18946 [cs]).
Abstract | Links | BibTeX | Tags:
@inproceedings{chemingui_constraint-adaptive_2024,
title = {Constraint-Adaptive Policy Switching for Offline Safe Reinforcement Learning},
author = {Yassine Chemingui and Aryan Deshwal and Honghao Wei and Alan Fern and Janardhan Rao Doppa},
url = {http://arxiv.org/abs/2412.18946},
doi = {10.48550/arXiv.2412.18946},
year = {2024},
date = {2024-12-01},
urldate = {2024-12-01},
publisher = {arXiv},
abstract = {Offline safe reinforcement learning (OSRL) involves learning a decision-making policy to maximize rewards from a fixed batch of training data to satisfy pre-defined safety constraints. However, adapting to varying safety constraints during deployment without retraining remains an under-explored challenge. To address this challenge, we introduce constraint-adaptive policy switching (CAPS), a wrapper framework around existing offline RL algorithms. During training, CAPS uses offline data to learn multiple policies with a shared representation that optimize different reward and cost trade-offs. During testing, CAPS switches between those policies by selecting at each state the policy that maximizes future rewards among those that satisfy the current cost constraint. Our experiments on 38 tasks from the DSRL benchmark demonstrate that CAPS consistently outperforms existing methods, establishing a strong wrapper-based baseline for OSRL. The code is publicly available at https://github.com/yassineCh/CAPS.},
note = {arXiv:2412.18946 [cs]},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Offline safe reinforcement learning (OSRL) involves learning a decision-making policy to maximize rewards from a fixed batch of training data to satisfy pre-defined safety constraints. However, adapting to varying safety constraints during deployment without retraining remains an under-explored challenge. To address this challenge, we introduce constraint-adaptive policy switching (CAPS), a wrapper framework around existing offline RL algorithms. During training, CAPS uses offline data to learn multiple policies with a shared representation that optimize different reward and cost trade-offs. During testing, CAPS switches between those policies by selecting at each state the policy that maximizes future rewards among those that satisfy the current cost constraint. Our experiments on 38 tasks from the DSRL benchmark demonstrate that CAPS consistently outperforms existing methods, establishing a strong wrapper-based baseline for OSRL. The code is publicly available at https://github.com/yassineCh/CAPS.
9.
Lily Oliphant
Agricultural Innovations CDE: Preparing Students for Agriculture 4.0 Journal Article
In: NAAE The Agricultural Education Magazine, vol. 97, no. 3, pp. 43–45, 2024.
@article{oliphant_agricultural_2024,
title = {Agricultural Innovations CDE: Preparing Students for Agriculture 4.0},
author = {Lily Oliphant},
url = {https://www.naae.org/naae/document-server/?cfp=/naae/assets/file/public/magazine/volume%2097/aged%20magazine_novemberdecember%202024%20-%20final.pdf},
year = {2024},
date = {2024-12-01},
urldate = {2024-12-01},
journal = {NAAE The Agricultural Education Magazine},
volume = {97},
number = {3},
pages = {43\textendash45},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
10.
Ranjan Sapkota; Zhichao Meng; Manoj Karkee
Synthetic meets authentic: Leveraging LLM generated datasets for YOLO11 and YOLOv10-based apple detection through machine vision sensors Journal Article
In: Smart Agricultural Technology, vol. 9, pp. 100614, 2024, ISSN: 2772-3755.
Abstract | Links | BibTeX | Tags:
@article{sapkota_synthetic_2024,
title = {Synthetic meets authentic: Leveraging LLM generated datasets for YOLO11 and YOLOv10-based apple detection through machine vision sensors},
author = {Ranjan Sapkota and Zhichao Meng and Manoj Karkee},
url = {https://www.sciencedirect.com/science/article/pii/S2772375524002193},
doi = {10.1016/j.atech.2024.100614},
issn = {2772-3755},
year = {2024},
date = {2024-12-01},
urldate = {2024-12-01},
journal = {Smart Agricultural Technology},
volume = {9},
pages = {100614},
abstract = {Training machine learning (ML) models for artificial intelligence (AI) and computer vision-based object detection process typically requires large, labeled datasets, a process often burdened by significant human effort and high costs associated with imaging systems and image acquisition. This research aimed to simplify image data collection for object detection in orchards by avoiding traditional fieldwork with different imaging sensors. Utilizing OpenAI's DALLE, a large language model (LLM) for realistic image generation, we generated and annotated a cost-effective dataset. This dataset, exclusively generated by LLM, was then utilized to train two state-of-the-art deep learning models: YOLOV10 and YOLO11. The YOLO11 model for apple detection was trained with its five configurations (YOLO11n, YOLO11 s, YOLO11 m, YOLO11l and YOLO11x), and YOLOv10 model with its six configurations (YOLOv10n, YOLOv10 s, YOLOv10 m, YOLOv10b, YOLOv10l and YOLOv10x), which was then tested with real-world (outdoor orchard) images captured by a digital (Nikon D5100) camera and a consumer RGB-D camera (Microsoft Azure Kinect). YOLO11 outperformed YOLOv10 as YOLO11x and YOLO11n exhibited superior precision of 0.917 and 0.916, respectively. Furthermore, YOLO11l demonstrated the highest recall among its counterparts, achieving a recall of 0.889. Likewise, the YOLO11n variant excelled in terms of mean average precision (mAP@50), achieving the highest value of 0.958. Validation tests against actual images collected through a digital camera (Nikon D5100) over Scilate apple variety in a commercial orchard environment showed a highest precision of 0.874 for YOLO11 s, recall of 0.877 for YOLO11l and mAP@50 of 0.91 for YOLO11x. Additionally, validation test against actual images collected through a Microsoft Azure camera over the same orchard showed a highest precision, recall and mAP@50 respectively of 0.924, 0.781 and 0.855 with YOLO11x. All variants of YOLO11 surprisingly demonstrated a pre-processing time of just 0.2 milliseconds (ms), which was faster than any variant of YOLOv10. The fastest inference time for the YOLO11n model using the training dataset generated by the language model was 3.2 ms, while YOLOv10n, fastest among YOLOv10 variants, had a longer inference time of 5.5 ms. Likewise, the fastest inference time for the sensor-based images was 7.1 ms (for Nikon D5100 camera images) and 4.7 ms (for Azure images) with YOLO11n. This study presents a pathway for generating large image datasets using LLM in challenging agricultural fields with minimal or no labor-intensive efforts in field data-collection, which could accelerate the development and deployment of computer vision and robotic technologies in orchard environments.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Training machine learning (ML) models for artificial intelligence (AI) and computer vision-based object detection process typically requires large, labeled datasets, a process often burdened by significant human effort and high costs associated with imaging systems and image acquisition. This research aimed to simplify image data collection for object detection in orchards by avoiding traditional fieldwork with different imaging sensors. Utilizing OpenAI's DALLE, a large language model (LLM) for realistic image generation, we generated and annotated a cost-effective dataset. This dataset, exclusively generated by LLM, was then utilized to train two state-of-the-art deep learning models: YOLOV10 and YOLO11. The YOLO11 model for apple detection was trained with its five configurations (YOLO11n, YOLO11 s, YOLO11 m, YOLO11l and YOLO11x), and YOLOv10 model with its six configurations (YOLOv10n, YOLOv10 s, YOLOv10 m, YOLOv10b, YOLOv10l and YOLOv10x), which was then tested with real-world (outdoor orchard) images captured by a digital (Nikon D5100) camera and a consumer RGB-D camera (Microsoft Azure Kinect). YOLO11 outperformed YOLOv10 as YOLO11x and YOLO11n exhibited superior precision of 0.917 and 0.916, respectively. Furthermore, YOLO11l demonstrated the highest recall among its counterparts, achieving a recall of 0.889. Likewise, the YOLO11n variant excelled in terms of mean average precision (mAP@50), achieving the highest value of 0.958. Validation tests against actual images collected through a digital camera (Nikon D5100) over Scilate apple variety in a commercial orchard environment showed a highest precision of 0.874 for YOLO11 s, recall of 0.877 for YOLO11l and mAP@50 of 0.91 for YOLO11x. Additionally, validation test against actual images collected through a Microsoft Azure camera over the same orchard showed a highest precision, recall and mAP@50 respectively of 0.924, 0.781 and 0.855 with YOLO11x. All variants of YOLO11 surprisingly demonstrated a pre-processing time of just 0.2 milliseconds (ms), which was faster than any variant of YOLOv10. The fastest inference time for the YOLO11n model using the training dataset generated by the language model was 3.2 ms, while YOLOv10n, fastest among YOLOv10 variants, had a longer inference time of 5.5 ms. Likewise, the fastest inference time for the sensor-based images was 7.1 ms (for Nikon D5100 camera images) and 4.7 ms (for Azure images) with YOLO11n. This study presents a pathway for generating large image datasets using LLM in challenging agricultural fields with minimal or no labor-intensive efforts in field data-collection, which could accelerate the development and deployment of computer vision and robotic technologies in orchard environments.
