Women in Engineering

Machine Learning is an indispensable part of data science, and there is no need to have a thorough programming background to benefit from it. Machine Learning (ML) and statistical techniques have provided a new era that enables us to convert the data into information and transform it into actionable knowledge. SciKit and TensorFlow are two states of the art libraries that can be used in Python, and this seminar will open the gate to know their bases. The first seminar is about “Hello World!” Machine Learning program, using the python language and SciKit learn library. Speaker(s): Dr. Reza Dibaj Virtual: https://events.vtools.ieee.org/m/276938

Energy efficiency, which has emerged as a top priority in cloud ecosystems, is the outcome of appropriate pricing mechanisms and resource allocations. Static pricing mechanisms are the most dominant approach among all the others. They are simple to implement for the service providers and easy to understand for the service users. Inaccurate price calculation and low efficient resource allocation in static pricing mechanisms made researchers discover other solutions to overcome these issues. Double auction mechanisms are among the most appropriate dynamic models. The main challenge of conventional double auction mechanisms is not considering the cloud ecosystems' specifications, such as dynamic online features. The term dynamic refers to the many variable parameters in cloud ecosystems, and they constantly change. Conventional static offline pricing solutions are set based on a series of parameters before running the process. In dynamic online methods, we customize our pricing models based on dynamic and current parameters. Also, we continuously optimize these methods to attain optimal results. In this seminar, firstly, we define a Dynamic Online Double Auction Mechanism (DODAM) for the IaaS environment, which covers a broader range of IaaS parameters by considering the dynamic online features of such markets. Considering the features of cloud dynamic online ecosystems, DODAM provides an appropriate price scheduling for service providers and service users. Cloud secondary market is a new paradigm in IaaS ecosystems. In these markets, brokers and reseller buyers have attained their resources from service providers of the cloud primary markets in the form of timed packages and repackage them into smaller chunks. As unsold packages do not transfer to the next intervals, brokers and reseller buyers need to sell their packages as much as possible. We develop a mechanical design that includes a market-based pricing model and a resource allocation algorithm in such markets as our second contribution. Next, by formulating the inherent competitive features in cloud secondary markets, we improve the pricing and resource allocation mechanisms in such competitive ecosystems. In the last contribution, we proposed a Priority-based Dynamic Online Double Auction Model (PB-DODAM), considering the perishability and time constraints of traded resources in IaaS secondary markets. The provided experimental results show that all proposed mechanisms drastically increase resource utilization and the overall utility. Speaker(s): Dr. Reza Dibaj Virtual: https://events.vtools.ieee.org/m/276942