scholarly journals A Sparse Deep Transfer Learning Model and Its Application for Smart Agriculture

2021 ◽  
Vol 2021 ◽  
pp. 1-11
Author(s):  
Zhikui Chen ◽  
Xu Zhang ◽  
Shi Chen ◽  
Fangming Zhong
Keyword(s):  
Transfer Learning ◽  
Expert Knowledge ◽  
Simplified Model ◽  
Power Devices ◽  
Final Model ◽  
Computing Power ◽  
Structural Differences ◽  
Sparse Networks ◽  
Smart Agriculture ◽  
Few Data

The introduction of deep transfer learning (DTL) further reduces the requirement of data and expert knowledge in various uses of applications, helping DNN-based models effectively reuse information. However, it often transfers all parameters from the source network that might be useful to the task. The redundant trainable parameters restrict DTL in low-computing-power devices and edge computing, while small effective networks with fewer parameters have difficulty transferring knowledge due to structural differences in design. For the challenge of how to transfer a simplified model from a complex network, in this paper, an algorithm is proposed to realize a sparse DTL, which only transfers and retains the most necessary structure to reduce the parameters of the final model. Sparse transfer hypothesis is introduced, in which a compressing strategy is designed to construct deep sparse networks that distill useful information in the auxiliary domain, improving the transfer efficiency. The proposed method is evaluated on representative datasets and applied for smart agriculture to train deep identification models that can effectively detect new pests using few data samples.

scholarly journals Transfer Learning of a Deep Learning Model for Exploring Tourists’ Urban Image Using Geotagged Photos

2021 ◽  
Vol 10 (3) ◽  
pp. 137
Author(s):  
Youngok Kang ◽  
Nahye Cho ◽  
Jiyoung Yoon ◽  
Soyeon Park ◽  
Jiyeon Kim
Keyword(s):  
Deep Learning ◽  
Transfer Learning ◽  
Field Studies ◽  
Learning Model ◽  
Learning Technologies ◽  
Regional Study ◽  
Processing Technologies ◽  
Final Model ◽  
Urban Image ◽  
Deep Learning Model

Recently, as computer vision and image processing technologies have rapidly advanced in the artificial intelligence (AI) field, deep learning technologies have been applied in the field of urban and regional study through transfer learning. In the tourism field, studies are emerging to analyze the tourists’ urban image by identifying the visual content of photos. However, previous studies have limitations in properly reflecting unique landscape, cultural characteristics, and traditional elements of the region that are prominent in tourism. With the purpose of going beyond these limitations of previous studies, we crawled 168,216 Flickr photos, created 75 scenes and 13 categories as a tourist’ photo classification by analyzing the characteristics of photos posted by tourists and developed a deep learning model by continuously re-training the Inception-v3 model. The final model shows high accuracy of 85.77% for the Top 1 and 95.69% for the Top 5. The final model was applied to the entire dataset to analyze the regions of attraction and the tourists’ urban image in Seoul. We found that tourists feel attracted to Seoul where the modern features such as skyscrapers and uniquely designed architectures and traditional features such as palaces and cultural elements are mixed together in the city. This work demonstrates a tourist photo classification suitable for local characteristics and the process of re-training a deep learning model to effectively classify a large volume of tourists’ photos.

EMPIRICAL FUZZY ESTIMATE OF THE SURVIVAL CURVE

2004 ◽  
Vol 12 (03) ◽  
pp. 347-356
Author(s):  
KIAN POKORNY ◽  
DILEEP SULE
Keyword(s):  
Expert Knowledge ◽  
Survival Curve ◽  
Complete Data ◽  
Point Estimate ◽  
Data Set ◽  
Fuzzy Membership Functions ◽  
Product Limit Estimator ◽  
Product Limit ◽  
Few Data ◽  
Fuzzy Estimate

In this paper, a computational system is developed that estimates a survival curve and a point estimate when very few data are available and a high proportion of the data are censored. Standard statistical methods require a more complete data set. With any less data expert knowledge or heuristic methods are required. The system uses numerical methods to define fuzzy membership functions about each data point that quantify uncertainty due to censoring. The "fuzzy" data is then used to estimate a survival curve and the mean survival time is calculated from the curve. The new estimator converges to the Product-Limit estimator when a complete data set is available. In addition, this method allows for the incorporation of expert knowledge. Finally, simulation results are provided to demonstrate the performance of the new method and its improvement over the Product-Limit estimator.

scholarly journals Impact of Internet of Things (IoT) in Smart Agriculture

2021 ◽  
Author(s):  
O. Vishali Priya ◽  
R. Sudha
Keyword(s):  
Internet Of Things ◽  
The Internet ◽  
Agricultural Field ◽  
Knowledge Processing ◽  
Computing Power ◽  
Crop Monitoring ◽  
The World ◽  
Instruments And Techniques ◽  
Smart Agriculture ◽  
The Internet Of Things

In today’s world, technology is constantly evolving; various instruments and techniques are available in the agricultural field. And within the agrarian division, the IoT preferences are Knowledge processing. With the help of introduced sensors, all information can be gathered. The reduction of risks, the mechanization of industry, the enhancement of production, the inspection of livestock, the monitoring of environment conditions, the roboticization of greenhouses, and crop monitoring Nearly every sector, like smart agriculture, has been modified by Internet-of-Things (IoT)-based technology, which has shifted the industry from factual to quantitative approaches. The ideas help to link real devices that are equipped with sensors, actuators, and computing power, allowing them to collaborate on a task while staying connected to the Internet, dubbed the “Internet of Things” (IoT). According to the World Telecommunication Union’s Worldwide Guidelines Operation, the Internet of Things (IoT) is a set of sensors, computers, software, and other devices that are connected to the Internet. The paper is highly susceptible to the consequences of its smart agriculture breakthrough.

scholarly journals Predictive model for in-hospital mortality following cardiogenic shock

2021 ◽  
Vol 42 (Supplement_1) ◽  
Author(s):  
F Garcia-Rodeja Arias ◽  
M Perez Dominguez ◽  
J Martinon Martinez ◽  
J M Garcia Acuna ◽  
C Abou Joch Casas ◽  
...  
Keyword(s):  
Hospital Mortality ◽  
Cardiogenic Shock ◽  
Predictive Model ◽  
Risk Score ◽  
Roc Curve ◽  
Goodness Of Fit ◽  
Simplified Model ◽  
P Value ◽  
Final Model ◽  
Risk Of Death

Abstract Introduction and objectives Cardiogenic shock is a condition caused by reduced cardiac output and hypotension, resulting in end-organ damage and multiorgan failure. Although prognosis has been improved in recent years, this state is still associated with high morbidity and mortality. The aim of our study was to perform a predictive model for in-hospital mortality that allows stratifying the risk of death in patients with cardiogenic shock. Methods This is a retrospective analysis from a prospective registry, that included 135 patients from one Spanish Universitary Hospital between 2011 and 2020. Multivariate analysis was performed among those variables with significant association with short-term outcome of univariate analysis with a p-value <0.2. Those variables which had a p-value >0.1 in the multivariable analysis were excluded of the final model. Our method was assessed using the area under the ROC-curve (AUC). Goodness of fit was tested using Hosmer-Lemeshow statistic test. Finally, we performed a risk score using the pondered weight of the coefficients of a simplified model created after categorizing the continuous quantitative variables included in the final model, giving a maximum of 16 points and creating three categories of risk. Results The in-hospital mortality rate was 41.5%, the average of age was 74.2 years, 35.6% were females and acute coronary syndrome (ACS) was the main cause of shock (60.7%). Mitral regurgitation (moderate-severe), age, ACS etiology, NT-proBNP, blood hemoglobin and lactate at admission were included in the final model. Risk-adjustment model had good accuracy in predicting in-hospital mortality (AUC 0.85; 95% CI 0,78–0,90) and the goodness of fit test was p-value>0.10. According to the risk score made with the simplified model, these patients were stratified into three categories: low (scores 0–6), intermediate (scores 7–10), and high (scores 11–16) risk with observed mortality of 12.9%, 49.1% and 87.5% respectively (p<0,001). Conclusions Our predictive model using six variables, shows good discernment for in-hospital mortality and the risk score has identified three groups with significant differences in prognosis. This model could help in guiding treatments and clinical decision-making, so it needs external validation and to be compared with other models already published. FUNDunding Acknowledgement Type of funding sources: None. ROC curve Risk Score

Urgently strategic insights to resolve the Nigerian food security crisis

2019 ◽  
Vol 49 (1) ◽  
pp. 77-85 ◽  
Author(s):  
Kenneth David Strang ◽  
Ferdinand Che ◽  
Narasimha Rao Vajjhala
Keyword(s):  
Developing Countries ◽  
Food Security ◽  
Expert Knowledge ◽  
Theoretical Models ◽  
Consensual Qualitative Research ◽  
Specialized Knowledge ◽  
Final Model ◽  
Local Agriculture ◽  
Rural Nigeria ◽  
Agriculture Extension

The food security crisis is a serious worldwide predicament in developing countries but it is a relatively larger problem in Nigeria. We argued there was no solution for the Nigerian food security crisis because researchers had not customized theoretical models with data-driven priorities grounded on local agriculture subject matter expert knowledge. We collected data from local agriculture extension workers who had specialized knowledge of the problems. We applied the consensual qualitative research method with embedded nominal brainstorming and multiple correspondence statistical techniques at the group level of analysis to develop a proposed solution. Our final model highlighted strategically urgent ideas to increase agriculture productivity and appease the most severe constraints in rural Nigeria. The results extended what was already published in the literature and should generalize to rural farmers in Nigeria as well as to government policymakers in developing countries around the world.

TLDS: A Transfer-Learning-Based Delivery Station Location Selection Pipeline

2021 ◽  
Vol 12 (4) ◽  
pp. 1-24
Author(s):  
Chenyu Hou ◽  
Bin Cao ◽  
Sijie Ruan ◽  
Jing Fan
Keyword(s):  
Transfer Learning ◽  
Expert Knowledge ◽  
Location Selection ◽  
Logistics Systems ◽  
Selection For ◽  
New City ◽  
Real World Datasets ◽  
Station Location ◽  
Delivery Efficiency ◽  
Range Selection

Delivery stations play important roles in logistics systems. Well-designed delivery station planning can improve delivery efficiency significantly. However, existing delivery station locations are decided by experts, which requires much preliminary research and data collection work. It is not only time consuming but also expensive for logistics companies. Therefore, in this article, we propose a data-driven pipeline that can transfer expert knowledge among cities and automatically allocate delivery stations. Based on existing well-designed station location planning in the source city, we first train a model to learn the expert knowledge about delivery range selection for each station. Then we transfer the learned knowledge to a new city and design three strategies to select delivery stations for the new city. Due to the differences in characteristics among different cities, we adopt a transfer learning method to eliminate the domain difference so that the model can be adapted to a new city well. Finally, we conduct extensive experiments based on real-world datasets and find the proposed method can solve the problem well.

Sign in / Sign up

Export Citation Format

Share Document