Khaled Rasheed

Previous work used machine learning (ML) to estimate past and current alfalfa yields and showed that domain adaptation (DA) with data synthesis shows promise in classifying yields as high, medium, or low. The current work uses similar techniques to forecast future alfalfa yields. A novel technique is proposed for forecasting alfalfa time series data that exploits stationarity and predicts differences in yields rather than the yields themselves. This forecasting technique generally provides more accurate forecasts than the established ARIMA family of forecasters for both univariate and multivariate time series. Furthermore, this ML-based technique is potentially easier to use than the ARIMA family of models. Also, previous work is extended by showing that DA with data synthesis also works well for predicting continuous values, not just for classification. The novel scale-invariant tabular synthesizer (SITS) is proposed, and it is competitive with or superior to other established synthesizers in producing data that trains strong models. This synthesis algorithm leads to R scores over 100% higher than an established synthesizer in this domain, while ML-based forecasters beat the ARIMA family with symmetric mean absolute percent error (sMAPE) scores as low as 12.81%. Finally, ML-based forecasting is combined with DA (ForDA) to create a novel pipeline that improves forecast accuracy with sMAPE scores as low as 9.81%. As alfalfa is crucial to the global food supply, and as climate change creates challenges with managing alfalfa, this work hopes to help address those challenges and contribute to the field of ML.

This study provides a comprehensive survey of the top-performing research papers in the field of time series prediction, offering insights into the most effective machine learning techniques, including tree-based, deep learning, and hybrid methods. It explores key factors influencing the model performance, such as the type of time series task, dataset size, and the time interval of historical data. Additionally, this study investigates potential biases in model development and weighs the trade-offs between the computational costs and performance. A detailed analysis of the most used error metrics and hyperparameter tuning methods in the reviewed papers is included. Furthermore, this study evaluates the results from prominent forecasting competitions, such as M5 and M6, to enrich the analysis. The findings of this paper highlight that tree-based methods like LightGBM 4.6.0 and deep learning methods like recurrent neural networks deliver the best performance in time series forecasting, with tree-based methods offering a significant advantage in terms of their computational efficiency. This paper concludes with practical recommendations for approaching time series forecasting tasks, offering valuable insights and actionable strategies that can enhance the accuracy and reliability of predictions derived from time series data.

Machine Learning (ML) models were used to determine the percent moisture content (MC) for multiple grains and seeds after training on a large dataset obtained through several decades of research. The dataset consisted of attenuation, phase shift, dielectric properties, frequency, bulk density, and sample thickness collected for corn, barley, sorghum, soybeans, and wheat. In this paper, a new ML-based approach for calibrating microwave sensors for rapid and nondestructive determination of moisture content in multiple grains and seeds is proposed. For this purpose, a single model trained on multiple grains and seeds was developed and allowed moisture determination in individual grain or seed samples. Performance of this model is investigated and compared with models trained on an individual grain or seed by using different algorithms including artificial neural network (NN), Support Vector Regression (SVR), ElasticNet, among other algorithms. In addition, these models were tested on new data collected for corn, wheat, and soybeans at 24∘ C with MC ranging from 7.89% to 20.19% and frequencies between 5 GHz to 15 GHz. The lowest mean absolute error (MAE) of MC was obtained with frequencies between 8 and 12 GHz for most models. Training when using the dielectric properties, frequency, and grain type with a single SVR-based model had the lowest error at 9 GHz for soybeans and corn. The SVR-based model showed no drawbacks and a slight improvement predicting MC using a single model when training over all grains and seeds compared with training several models over each grain individually.

Deep learning has been the answer to many machine learning problems during the past two decades. However, it comes with two significant constraints: dependency on extensive labeled data and training costs. Transfer learning in deep learning, known as Deep Transfer Learning (DTL), attempts to reduce such reliance and costs by reusing obtained knowledge from a source data/task in training on a target data/task. Most applied DTL techniques are network/model-based approaches. These methods reduce the dependency of deep learning models on extensive training data and drastically decrease training costs. Moreover, the training cost reduction makes DTL viable on edge devices with limited resources. Like any new advancement, DTL methods have their own limitations, and a successful transfer depends on specific adjustments and strategies for different scenarios. This paper reviews the concept, definition, and taxonomy of deep transfer learning and well-known methods. It investigates the DTL approaches by reviewing applied DTL techniques in the past five years and a couple of experimental analyses of DTLs to discover the best practice for using DTL in different scenarios. Moreover, the limitations of DTLs (catastrophic forgetting dilemma and overly biased pre-trained models) are discussed, along with possible solutions and research trends.

Deep learning (DL) along with never-ending advancements in computational processing and cloud technologies have bestowed us powerful analyzing tools and techniques in the past decade and enabled us to use and apply them in various fields of study. Health informatics is not an exception, and conversely, is the discipline that generates the most amount of data in today's era and can benefit from DL the most. Extracting features and finding complex patterns from a huge amount of raw data and transforming them into knowledge is a challenging task. Besides, various DL architectures have been proposed by researchers throughout the years to tackle different problems. In this paper, we provide a review of DL models and their broad application in bioinformatics and healthcare categorized by their architecture. In addition, we also go over some of the key challenges that still exist and can show up while conducting DL research.

Alfalfa is critical to global food security, and its data is abundant in the U.S. nationally, but often scarce locally, limiting the potential performance of machine learning (ML) models in predicting alfalfa biomass yields. Training ML models on local-only data results in very low estimation accuracy when the datasets are very small. Therefore, we explore synthesizing non-local data to estimate biomass yields labeled as high, medium, or low. One option to remedy scarce local data is to train models using non-local data; however, this only works about as well as using local data. Therefore, we propose a novel pipeline that trains models using data synthesized from non-local data to estimate local crop yields. Our pipeline, synthesized non-local training (SNLT pronounced like sunlight), achieves a gain of 42.9% accuracy over the best results from regular non-local and local training on our very small target dataset. This pipeline produced the highest accuracy of 85.7% with a decision tree classifier. From these results, we conclude that SNLT can be a useful tool in helping to estimate crop yields with ML. Furthermore, we propose a software application called Predict Your CropS (PYCS pronounced like Pisces) designed to help farmers and researchers estimate and predict crop yields based on pretrained models.

Deep transfer learning techniques try to tackle the limitations of deep learning, the dependency on extensive training data and the training costs, by reusing obtained knowledge. However, the current DTL techniques suffer from either catastrophic forgetting dilemma (losing the previously obtained knowledge) or overly biased pre-trained models (harder to adapt to target data) in finetuning pre-trained models or freezing a part of the pre-trained model, respectively. Progressive learning, a sub-category of DTL, reduces the effect of the overly biased model in the case of freezing earlier layers by adding a new layer to the end of a frozen pre-trained model. Even though it has been successful in many cases, it cannot yet handle distant source and target data. We propose a new continual/progressive learning approach for deep transfer learning to tackle these limitations. To avoid both catastrophic forgetting and overly biased-model problems, we expand the pre-trained model by expanding pre-trained layers (adding new nodes to each layer) in the model instead of only adding new layers. Hence the method is named EXPANSE. Our experimental results confirm that we can tackle distant source and target data using this technique. At the same time, the final model is still valid on the source data, achieving a promising deep continual learning approach. Moreover, we offer a new way of training deep learning models inspired by the human education system. We termed this two-step training: learning basics first, then adding complexities and uncertainties. The evaluation implies that the two-step training extracts more meaningful features and a finer basin on the error surface since it can achieve better accuracy in comparison to regular training. EXPANSE (model expansion and two-step training) is a systematic continual learning approach applicable to different problems and DL models.

The total boll count from a plant is one of the most important phenotypic traits for cotton breeding and is also an important factor for growers to estimate the final yield. With the recent advances in deep learning, many supervised learning approaches have been implemented to perform phenotypic trait measurement from images for various crops, but few studies have been conducted to count cotton bolls from field images. Supervised learning models require a vast number of annotated images for training, which has become a bottleneck for machine learning model development. The goal of this study is to develop both fully supervised and weakly supervised deep learning models to segment and count cotton bolls from proximal imagery. A total of 290 RGB images of cotton plants from both potted (indoor and outdoor) and in-field settings were taken by consumer-grade cameras and the raw images were divided into 4350 image tiles for further model training and testing. Two supervised models (Mask R-CNN and S-Count) and two weakly supervised approaches (WS-Count and CountSeg) were compared in terms of boll count accuracy and annotation costs. The results revealed that the weakly supervised counting approaches performed well with RMSE values of 1.826 and 1.284 for WS-Count and CountSeg, respectively, whereas the fully supervised models achieve RMSE values of 1.181 and 1.175 for S-Count and Mask R-CNN, respectively, when the number of bolls in an image patch is less than 10. In terms of data annotation costs, the weakly supervised approaches were at least 10 times more cost efficient than the supervised approach for boll counting. In the future, the deep learning models developed in this study can be extended to other plant organs, such as main stalks, nodes, and primary and secondary branches. Both the supervised and weakly supervised deep learning models for boll counting with low-cost RGB images can be used by cotton breeders, physiologists, and growers alike to improve crop breeding and yield estimation.