The increasingly growing expansion of the Internet of Things (IoT) along with the convergence of multiple technologies such as the arrival of next generation wireless broadband in 5G, is creating a paradigm shift from cloud computing towards edge computing. Performing tasks normally done by the cloud directly on edge devices would ensure multiple benefits such as latency gains and a more robust privacy of data. However, edge devices are resource-constrained and often do not possess the computational and memory capabilities to perform demanding tasks. Complex algorithms such as the training and inference of a complete Deep Neural Network (DNN) is often not feasible on these devices. In this paper we perform a novel empirical study of the various ways that multiple inference tasks of deep learning models can be loaded on these edge devices. We analyse the run time gain, under different resource limits, of various DNN layer loading policies that aim to optimize the overall run time of consecutive inference tasks. We combine this with further research in the memory usage and swapping behaviour when performing these inference tasks. Using these results, we show that if the memory overhead becomes too large, loading and executing DNN layers in an interleaved manner provides significant gains in run time. This is achieved trough multiple experiments in our specially made evaluation environment EdgeCaffe which is presented in this paper as well.

The most crucial choices a student will make is about which college and major they decide to join. Accord- ing to a statistical analysis performed by Koenig (2018) in the U.S. News World Report, majors such as Computer and information science, Engineering and Engineering technology yield the highest employment rates and salaries compared to other majors. In an article they wrote about the factors that influence youths career choices, Akosah-Twumasi (2018) argued that the knowledge of issues related to ’job security’ and ’salaries’ may pressure youth to choose a career path based on the benefits associated with a particular profession. This causes an influence in the decision making of a student who will not necessarily apply for a major they would enjoy doing, but instead their choice is going to shift to a more reliable major. Thus, many students will apply for studies such as Computer Science even though it might not be well-suited for them. Our team has been asked by the Delft University of Technology’s communication department to develop a Chatbot in order to help students with their decision making, and specifically students interested in the master program Embedded Systems. The communication department gave our team a set of requirements that needed to be fulfilled. The final product needed to be a chatbot with which it is possible to have a conversation on the Embedded Systems study program. It should coach the student into making a decision as well as be able to answer frequently asked questions. The chatbot needed to be accompanied by a content management system which should allow the communication department to modify some of the content of the chatbot as well as provide them with useful statistics about the interactions with the chatbot. Our team was also required to use the Rasa (2019) open source machine learning tool for conversational artificial intelligence as back-end of our chatbot system in order to provide feedback about this framework which might be used in future projects at TU Delft.