Titulo Estágio
Try to beat the system
Áreas de especialidade
Engenharia de Software
Sistemas Inteligentes
Local do Estágio
Laboratórios do CISUC, DEI
Enquadramento
The importance of mobile devices in our lives makes it hard to imagine our daily activities without them. The use of smartphones, tablet computers, and, more recently, wearables such as smartwatches have changed and simplified the way we communicate with each other, how we individually or collectively have fun, and how we work or do business. Ultimately, the number and scope of mobile applications seem unlimited, with users expecting more and more from them each day. The distribution of such applications is highly facilitated by app stores such as Aptoide, which democratize the opportunity to commercialize software for mobile devices. In 2017, the number of downloaded mobile applications was estimated at 178 billion, and this number is expected to grow to 258 billion by 2022.
While the app market targets mobile devices, which by nature most often run on batteries, the fact is that current marketplaces do not provide any indication of the absolute or relative energy efficiency of the applications they host. Having this breach in mind, on GreenStamp, we propose to research and develop innovative mechanisms of analysis and profiling of the energy efficiency of the apps in app stores. Based on this analysis, recommendation systems for app developers will be studied, how to improve their apps' energy efficiency, and to users about the energy efficiency of the apps they are about to install, which may influence the autonomy of their devices. As the main objective, we have the reduction of 20% of the energy consumption by apps that follow the technical recommendations provided and, inherently, of the devices where they are installed, contributing thus to significant savings on the energy consumed by Aptoide's user base (220 million unique active users in 2018).
This internship is eligible for a scholarship.
Objetivo
The aim of Greenstamp is to develop a sound approach for classifying mobile apps accordingly to their energy consumption profile. This is like a benchmark, which will enable developers, providers, and users to compare apps concerning their power consumption behavior. When a benchmark becomes a reference for a target market, providers begin to design and develop their products to (artificially) rank as high as possible with the benchmark. Thus, the benchmark slowly becomes less relevant since it may no longer be representative of reality.
This internship aims to create adversarial strategies targeted to evaluate Greestamp that would make an Apps rank higher than it should, testing its robustness and analyzing potential exploits of the system.
Plano de Trabalhos - Semestre 1
Step 1 – Definition of the research question(s) and project scope
Step 2 – Review the literature and writing of the State-of-the-Art
Step 3 – Definition of the methodology and approach
Step 4 - Implementation of a proof-of-concept
Step 5 - Writing the intermediate report
Plano de Trabalhos - Semestre 2
Step 6 – Setup of the experimental environment
Step 7 – Implementation and evaluation of the proposed adversarial strategies
Step 8 - Writing the final report
Condições
The work will be conducted in the Systems and Software Engineering group from CISUC. The eligible student will have all the necessary computational platforms, tools, and devices at their disposal.
There is a possibility of the student being awarded a scholarship. The scholarship will follow the Fundação para a Ciência e Tecnologia (FCT) monthly stipend guidelines (Bolsa de Investigação para Licenciado).
Observações
Co-advisor: Prof. João Correia, DEI-FCTUC
Orientador
Bruno Cabral
bcabral@dei.uc.pt 📩