scholarly journals Testing the Suitability of Automated Machine Learning for Weeds Identification

AI ◽  
2021 ◽  
Vol 2 (1) ◽  
pp. 34-47
Author(s):  
Borja Espejo-Garcia ◽  
Ioannis Malounas ◽  
Eleanna Vali ◽  
Spyros Fountas
Keyword(s):  
Machine Learning ◽  
Plant Protection ◽  
Crop Protection ◽  
Identification Problem ◽  
Learning System ◽  
Classifier Ensembles ◽  
Automated Machine Learning ◽  
A New Technique ◽  
Plant Seedlings ◽  
And Training

In the past years, several machine-learning-based techniques have arisen for providing effective crop protection. For instance, deep neural networks have been used to identify different types of weeds under different real-world conditions. However, these techniques usually require extensive involvement of experts working iteratively in the development of the most suitable machine learning system. To support this task and save resources, a new technique called Automated Machine Learning has started being studied. In this work, a complete open-source Automated Machine Learning system was evaluated with two different datasets, (i) The Early Crop Weeds dataset and (ii) the Plant Seedlings dataset, covering the weeds identification problem. Different configurations, such as the use of plant segmentation, the use of classifier ensembles instead of Softmax and training with noisy data, have been compared. The results showed promising performances of 93.8% and 90.74% F1 score depending on the dataset used. These performances were aligned with other related works in AutoML, but they are far from machine-learning-based systems manually fine-tuned by human experts. From these results, it can be concluded that finding a balance between manual expert work and Automated Machine Learning will be an interesting path to work in order to increase the efficiency in plant protection.

scholarly journals Epidemiology and forecasting of insect-pests and diseases for value-added agro-advisory

MAUSAM ◽  
2021 ◽  
Vol 67 (1) ◽  
pp. 267-276
Author(s):  
AMRENDER KUMAR ◽  
A. K. JAIN ◽  
B. K. BHATTACHARYA ◽  
VINOD KUMAR ◽  
A. K. MISHRA ◽  
...  
Keyword(s):  
Machine Learning ◽  
Insect Pests ◽  
Plant Protection ◽  
Crop Protection ◽  
Value Added ◽  
Accurate Estimation ◽  
Learning Approach ◽  
Pests And Diseases ◽  
Functional Relationships ◽  
Machine Learning Approach

Models are means to capture, condense and organize knowledge. These are expressions, which represent relationship between various components of a system. A well-tested weather-based model can be an effective scientific tool for forewarning insect-pests and diseases in advance so that timely plant protection measures could be taken up. Various types of techniques have been developed for the purpose. The simplest technique forms the class of thumb rules, which are based on experience. Though these do not have much scientific background but are extensively used to provide quick forewarning of the menace. Another tool in practice is regression model that represents relationship between two or more variables so that one variable can be predicted from the other (s). Linear and non-linear regression models have been widely used in studying relationship of insect-pests and diseases with time and weather variables (as such or in some transformed forms). With the advent of computers more sophisticated techniques such as simulation modelling and machine learning approach such as decision tree induction algorithms, genetic algorithms, neural networks, rough sets, etc. have been explored. A number of simulation models have been developed all over the world for quantifying effects of various factors including weather on agriculture.  These may provide a good forecast but require detailed data base, which may not be available. Machine learning approach has recently received some attention. As opposed to traditional model-based methods, machine learning approach is self adaptive methods in that there are a few a priori assumptions about the models for problem(s) under study. This technique learns more from examples and captures subtle functional relationships among the data even if the underlying relationships are unknown or hard to describe.  This modelling approach with ability to learn from experience is very useful for many practical problems provided enough data are available. Remotely sensed data can provide useful information relating to area under the crop and also the condition thereof. It has certain advantages over land use statistics due to multi-spectral, synoptic and repetitive coverage. An attempt has been made for accurate estimation of area affected by insect-pests and diseases in crops along with accurate assessment of damage due to the same are possible for providing compensation to farmers. In this study, an Integrated Decision Support System (IDSS) for Crop Protection Services is also discussed.  

scholarly journals SIMON, an automated machine learning system reveals immune signatures of influenza vaccine responses

2019 ◽  
Author(s):  
Adriana Tomic ◽  
Ivan Tomic ◽  
Yael Rosenberg-Hasson ◽  
Cornelia L. Dekker ◽  
Holden T. Maecker ◽  
...  
Keyword(s):  
Machine Learning ◽  
Statistical Power ◽  
Missing Values ◽  
Learning System ◽  
T Cell Subsets ◽  
Vaccine Responses ◽  
Biological Insight ◽  
Speed Up ◽  
Automated Machine Learning ◽  
Incomplete Datasets

AbstractMachine learning holds considerable promise for understanding complex biological processes such as vaccine responses. Capturing interindividual variability is essential to increase the statistical power necessary for building more accurate predictive models. However, available approaches have difficulty coping with incomplete datasets which is often the case when combining studies. Additionally, there are hundreds of algorithms available and no simple way to find the optimal one. Here, we developed Sequential Iterative Modelling “OverNight” or SIMON, an automated machine learning system that compares results from 128 different algorithms and is particularly suitable for datasets containing many missing values. We applied SIMON to data from five clinical studies of seasonal influenza vaccination. The results reveal previously unrecognized CD4+ and CD8+ T cell subsets strongly associated with a robust antibody response to influenza antigens. These results demonstrate that SIMON can greatly speed up the choice of analysis modalities. Hence, it is a highly useful approach for data-driven hypothesis generation from disparate clinical datasets. Our strategy could be used to gain biological insight from ever-expanding heterogeneous datasets that are publicly available.

scholarly journals An evolutionary algorithm for automated machine learning focusing on classifier ensembles: An improved algorithm and extended results

2020 ◽  
Vol 805 ◽  
pp. 1-18
Author(s):  
João C. Xavier-Júnior ◽  
Alex A. Freitas ◽  
Teresa B. Ludermir ◽  
Antonino Feitosa-Neto ◽  
Cephas A.S. Barreto
Keyword(s):  
Machine Learning ◽  
Evolutionary Algorithm ◽  
Classifier Ensembles ◽  
Automated Machine Learning ◽  
Improved Algorithm

scholarly journals SIMON, an Automated Machine Learning System, Reveals Immune Signatures of Influenza Vaccine Responses

2019 ◽  
Vol 203 (3) ◽  
pp. 749-759 ◽  
Author(s):  
Adriana Tomic ◽  
Ivan Tomic ◽  
Yael Rosenberg-Hasson ◽  
Cornelia L. Dekker ◽  
Holden T. Maecker ◽  
...  
Keyword(s):  
Machine Learning ◽  
Influenza Vaccine ◽  
Learning System ◽  
Vaccine Responses ◽  
Immune Signatures ◽  
Automated Machine Learning

scholarly journals Towards Automated Semi-Supervised Learning

2019 ◽  
Vol 33 ◽  
pp. 4237-4244 ◽  
Author(s):  
Yu-Feng Li ◽  
Hai Wang ◽  
Tong Wei ◽  
Wei-Wei Tu
Keyword(s):  
Machine Learning ◽  
Supervised Learning ◽  
Semisupervised Learning ◽  
Learning System ◽  
Large Margin ◽  
Automated Learning ◽  
Machine Learning Model ◽  
Meta Learning ◽  
Performance Deterioration ◽  
Automated Machine Learning

Automated Machine Learning (AutoML) aims to build an appropriate machine learning model for any unseen dataset automatically, i.e., without human intervention. Great efforts have been devoted on AutoML while they typically focus on supervised learning. In many applications, however, semisupervised learning (SSL) are widespread and current AutoML systems could not well address SSL problems. In this paper, we propose to present an automated learning system for SSL (AUTO-SSL). First, meta-learning with enhanced meta-features is employed to quickly suggest some instantiations of the SSL techniques which are likely to perform quite well. Second, a large margin separation method is proposed to fine-tune the hyperparameters and more importantly, alleviate performance deterioration. The basic idea is that, if a certain hyperparameter owns a high quality, its predictive results on unlabeled data may have a large margin separation. Extensive empirical results over 200 cases demonstrate that our proposal on one side achieves highly competitive or better performance compared to the state-of-the-art AutoML system AUTO-SKLEARN and classical SSL techniques, on the other side unlike classical SSL techniques which often significantly degenerate performance, our proposal seldom suffers from such deficiency.

The Impact of Plant Protection Systems on the Productivity of Soybean Varieties under Irrigated Conditions

2019 ◽  
pp. 109-117
Keyword(s):  
Plant Protection ◽  
Crop Protection ◽  
Production Costs ◽  
Protection System ◽  
Average Weight ◽  
The South ◽  
Biological Origin ◽  
Chemical Protection ◽  
Protection Systems ◽  
The Impact

The application of preparations of biological origin in the protection system of soybean grown under conditions of intensive irrigated crop rotations conforms to the modern tendencies of science and production development. The use of them contributes to solving ecological, production and social-economic problems. The study presents the three-year research on the efficiency of systems protecting soybean from pests and diseases based on biological and chemical preparations. The research was conducted in typical soil and climate conditions of the South of Ukraine. Zonal agricultural methods and generally accepted research methodology were used. The purpose of the research was to create a soybean protection system based on preparations of biological origin, ensuring high productivity of high-quality products reducing a negative impact of the crop production on the environment. The study emphasizes that, under irrigated conditions of the South of Ukraine, the application of biological preparations has a positive impact on the indexes of growth, development and formation of the elements of soybean yield structure. There was an increase in the crop biological weight by 13.8 % and 22.1 % and the number of seeds per plant rose by 11.6 and 14.6 % as a consequence of eliminating harmful organisms with the plant protection systems. The larger ground mass was formed by medium-ripe varieties Danai and Svyatogor, on which the increase from protection measures was higher. Weight 1000 pcs. the seeds did not undergo significant changes. It is established that the larger seeds were formed by Danaya and Svyatogor varieties, in which the average weight is 1000 pcs. seeds were 142 and 136 g, respectively, while in the variety Diona this figure was 133 g. There was an increase in the height of the lowest pod when the total plant height rose. For medium-ripe varieties was characterized by a higher attachment of beans, where the highest values of this indicator acquired in the variety Svyatogor. The medium maturing soybean variety Danaia formed the maximum yield of 3.23 and 3.35 t/ha respectively, when biological and chemical protection systems were applied. The research establishes that the application of the bio-fungicide Psevdobakterin 2 (2.0 l/ha) in the crop protection system at the beginning of soybean flowering and the bio-fungicide Baktofit (2.5 l/ha) with the bio-insecticide Lepidotsid-BTU (10.0 l/ha) at the beginning of pod formation does not reduce the productivity of the soybean varieties under study considerably, when compared to the application of chemical preparations. The research determines that the soybean protection system under study ensures a decrease in the coefficient of soybean water uptake by 7.2-13.0 %, increasing the total water intake to an inconsiderable degree. Biologization of the soybean crop protection system leads to a reduction in production costs compared to the chemical protection system. Taking into account the needs for the collection of additional products, costs increase by an average of 1 thousand UAH/ha, while for chemical protection systems by 1.8 thousand UAH/ha. At the same time, the cost is reduced by 220-360 UAH/t and the profitability of growing crops is increased by 3.8-7.8 %. There has been a reduction in the burden of pesticides on the environment and the production of cleaner products. This indicates the prospect of using the biofungicides Pseudobacterin 2 and Bactophyte and the bioinsecticide Lepidocid-BTU on soybeans to protect plants from pests.

Bandit-Based Automated Machine Learning

2018 ◽  
Author(s):  
Silvia Cristina Nunes das Dores ◽  
Carlos Soares ◽  
Duncan Ruiz
Keyword(s):  
Machine Learning ◽  
Automated Machine Learning
Sign in / Sign up

Export Citation Format

Share Document