WSO2 (Sri Lanka)

The growing relevance of research on gender and social inclusion in agricultural research for development calls for systemic, transformative change processes. Transformative gender ambitions can stand at odds with personal biases and experiences that shape diverse understandings of gender, institutional values, structures and cultures that tend to reward technological quick-fix solutions, and other practical challenges to 'doing' gender on the ground. Very little is known about these challenges. How are these challenges navigated by (relatively small) teams of gender researchers, who are often caught between the demand for tangible fast gains on gender, and the intractable challenges of deeprooted and complex, intersectional gender inequalities? This was the focus of the CGIAR Research Program on Water, Land and Ecosystems (WLE) End of Program Reflection and Evaluation (EoPRE) to assess how gender and inclusion research is pursued, and the key barriers to knowing and doing gender in eight research projects. Adopting a reflexive, self-analytical feminist approach to evaluation, this EoPRE facilitated eight project teams, diverse and with an uneven focus on gender, to connect the dots between the processes of knowing and doing gender research. A key finding of this evaluation is that the need for change is foremost internal. We need to begin by fixing our personal biases and assumptions, and fixing institutional cultures, values and structures instead of just trying to fix things out there, including fixing poor and marginalized women. A key recommendation is to seek more regular and open conversations across researcher disciplines and hierarchies, and between CGIAR and external partners and stakeholders, including feminist grassroots actors and networks -on what works well (and does not) and why. This would allow us to grasp why we start with different meanings and conceptualizations of gender; how agile we are (or not) in adapting to changes on the ground; and how, through a culture of reflection and learning, we might shift pathways to more transformative change processes in a fast evolving and increasingly unequal world.

In the retail domain, estimating the sales before actual sales become known plays a key role in maintaining a successful business. This is due to the fact that most crucial decisions are bound to be based on these forecasts. Statistical sales forecasting models like ARIMA (Auto-Regressive Integrated Moving Average), can be identified as one of the most traditional and commonly used forecasting methodologies. Even though these models are capable of producing satisfactory forecasts for linear time series data they are not suitable for analyzing non-linear data. Therefore, machine learning models (such as Random Forest Regression, XGBoost) have been employed frequently as they were able to achieve better results using non-linear data. The recent research shows that deep learning models (e.g. recurrent neural networks) can provide higher accuracy in predictions compared to machine learning models due to their ability to persist information and identify temporal relationships. In this paper, we adopt a special variant of Long Short Term Memory (LSTM) network called LSTM model with peephole connections for sales prediction. We first build our model using historical features for sales forecasting. We compare the results of this initial LSTM model with multiple machine learning models, namely, the Extreme Gradient Boosting model (XGB) and Random Forest Regressor model(RFR). We further improve the prediction accuracy of the initial model by incorporating features that describe the future that is known to us in the current moment, an approach that has not been explored in previous state-of-the-art LSTM based forecasting models. The initial LSTM model we develop outperforms the machine learning models achieving 12% - 14% improvement whereas the improved LSTM model achieves 11\% - 13\% improvement compared to the improved machine learning models. Furthermore, we also show that our improved LSTM model can obtain a 20% - 21% improvement compared to the initial LSTM model, achieving significant improvement.

Enterprise Service Bus (ESB) is a middleware that provides solutions for enterprise application integration. Although the contemporary ESB products exhibit diverse architectural styles and standards such as service component architecture and Java business integration, they mostly provide the same set of ESB services such as data transformation, security, et cetera. The quality attributes of a software system are primarily attributed to the system’s architecture, and a set of systems having different architectures can meet the requirements from a great variety of users. To produce several ESB variations successfully, a systematic reuse across ESB systems is crucial. Therefore, the commonality in ESB products, which is comprised mainly of ESB services, should be strategically exploited, and this chapter discusses an approach to realize it. The author presents a platform that can derive architecturally heterogeneous ESB products from reusable ESB services. The approach for building the platform leverages aspect-oriented programming.

The commonality across software systems can be exploited to develop multiple heterogeneous systems successfully without undue cost, time, and effort. The systematic reuse across different systems is of paramount importance. With a well-planned reuse approach, a vendor can offer individualized products, which are products tailored to meet the requirements of a particular user effectively, as well as the products constructed to deliver solutions for a greater variety of application domains such as enterprise application integration and business process management. This chapter describes the development of software systems having different architectures reusing most of the implementations of the required functionalities as-is. It presents a systematic process for crafting multi-architecture reusable components and for using those components in formulating software systems. Furthermore, the chapter highlights the significance of the strategic reuse across systems in three contemporary research spheres.

Enterprise Service Bus is an Enterprise Application Integration middleware. Although ESB products are heterogeneous in terms of architectures, most often, they provide the same set of ESB services such as message routing, security, etc. This emphasizes the fact that it should be possible to derive all ESB products from a single ESB service library, and we present an approach to realize it. In this paper, we present two main ingredients of our solution, namely an ESB service library and a service assembler. The ESB service library consists with ESB services, which are adaptable for most ESB architectures and designed to provide quality requirements such as performance, reusability, extensibility, etc. The service assembler enables a software architect to use most ESB architectures “as is” or further enhance the quality they already provide. Our approach for building the service assembler leverages Aspect Oriented Programming.