As a joint venture between Simula Research Laboratory and Oslo Metropolitan University (OsloMet), SimulaMet takes full advantage of the thriving, top level multicultural research culture of Simula, while being physically located at OsloMet, a large university with more than 20.000 students and 7 PhD programs.
SimulaMet is the home of Simula’s research activities on networks and communications, machine learning, and IT management. Its main objective is to generate new understanding and create vital knowledge about fundamental scientific challenges that are of genuine value for society. This is achieved through high quality research, education of graduate students, industry collaboration, technology transfer and commercialization.
PhD position in Machine Learning for Autonomous Intelligent Networks
SimulaMet: solving problems, making a difference.
“We hire the best people, those who are personally driven to push boundaries and innovate - and then we give them the tools, space and freedom to do just that.”
SIGIPRO at SimulaMet, is a vibrant interdisciplinary work environment that supports talented researchers to create innovative solutions that benefit society. Research quality is the foremost important criterion. We deliver innovative solutions for intelligent and multimodal sensor networks, information systems, and networked cyber-physical systems, by creating theories and algorithms that blend different disciplines. Our researchers work to establish synergies between fundamental theory, algorithmic designs and application-specific implementations. Our cross-disciplinary research approach combines tools and methods from signal processing, data science, machine learning, distributed cooperative intelligence, graph theory, numerical optimization, autonomous multi-agent systems, wireless networks, complexity analysis, and decision control theory. We investigate theoretical prescriptions that are used to guide the design of optimal (or close-to-optimal) efficient algorithms, and which provide performance guarantees when possible, so that any heuristics are properly justified. These algorithms are also tested both via simulation and using real platforms and testbed systems.
The PhD position will be associated to the project Distributed Learning and Cooperative Optimization for Multi-agent Autonomous Wireless Access Points (DISCO), funded by the Research Council of Norway, whose objective is to design a new generation of decentralized data-driven Machine Learning based methods for distributed learning and cooperative optimization over autonomous intelligent network ecosystems, with a special focus on large-scale autonomous cooperative wireless Access Points (APs).
Successful applicants will carry out research activities within one or both of the following areas:
Design of novel automated distributed online Machine Learning (ML) methods based on online decentralized optimization and graph signal processing, which can learn and track efficiently multi-modal dynamic graphs representing the global state of operation of complex dynamic systems, by using various multi-variate and multimodal data measurements. Examples of complex dynamic systems, among others, include multiple interfering (separately owned) Wi-Fi networks, e.g. in residential Wi-Fi deployments. These methods will infer which and how the relevant data variables are affecting each other and will also allow to perform forecasting of relevant variables, as well as missing data imputation and anomaly detection. For example, in the case of Wi-Fi networks, this dynamic graph will allow us to: a) detect clusters of APs such that each cluster will be composed of a set of interfering Wi-Fi networks, determining the set of APs whose operation parameters must be optimized to maximize performance metrics of interest, b) detect anomalies due to possible malfunctioning of APs or malicious behaviour.
Design of novel resource allocation schemes for next generation autonomous networks, by using distributed causality-enhanced Multi-Agent Reinforcement Learning (MARL) methods for learning-based adaptive control, which will exploit a state representation of the overall network dynamics. The multiple agents will interact with each other exchanging limited information, in order to perform adaptive sequential control of the various resource allocations, such as client to AP association, frequency band, channel and power allocation, through in-network cooperation across the APs, without communicating to a cloud center, and achieving a performance in terms of relevant metrics (e.g. throughput, latency, load balancing) that is superior to the currently existing solutions.
The main scientific methodologies will be based on: a) online distributed statistical learning methods (e.g. kernel-based, gaussian process regression) of causal dynamic graph structures to express dependencies among data variables, as well as various graph signal processing tools for multivariate data, b) distributed iterative optimization techniques for convex and non-convex problems, c) distributed MARL algorithms, under different possible architectures (e.g. distributed Actor-Critic algorithms), with the novelty of exploiting the learned causality graphs, with different possible decentralized aggregation and processing techniques, such as graph-time neural networks.
The candidate will join the SIGIPRO department, which includes several other highly motivated researchers, collaborating also directly with relevant national partners (e.g. Altibox, Multinett, UiO) and several other internationally recognized partners (e.g. Nokia Bell Labs, KTH, UPF), which participate also in the DISCO project. We offer you to work in a research environment with a unique combination of theoretical and practical experience targeting top international research quality, including also state-of-the-art laboratory resources, supercomputing capabilities, and advanced network testbeds.
PhD, online automated machine learning, statistical causal model learning, decentralized optimization, graph signal processing, intelligent agents, graph neural networks, learning-based control, multi-agent deep reinforcement learning, next generation autonomous wireless networks, 6G.
Hold a MSc degree in electrical engineering (or telecommunications), computer science, computer engineering, ICT, control engineering, or other related discipline, at the start date of the position.
Have work experience and/or research in data science and machine learning methods, especially in the areas of statistical model learning, numerical optimization regression techniques, graph neural networks, learning-based control, and reinforcement learning. Previous education, experience and/or research in wireless networks, control theory and autonomous systems, is a plus.
Have good programming skills, primarily in Python. Experience in Matlab, C/C++, container-based micro-services implementations, or other relevant tools (e.g. PyTorch, Reinforcement Learning environments, GitHub) is also a plus.
Have a strong interest in high quality scientific publications and dissemination.
Be highly motivated, enthusiastic, and be able to demonstrate their potential for conducting original interdisciplinary research, problem solving, having a critical and scientific thinking and being able to work both independently and collaboratively within a team.
Have good communication skills and be fluent in English, both writing and speaking.
Have good interpersonal skills to work with both academic and industrial stakeholders to support research and show willingness and ability to contribute to the SimulaMet dynamic and inclusive working environment.
Satisfy the enrolment requirements for the OsloMet PhD Program in Engineering Sciences. This includes: a) an average grade of C or better for the bachelor’s degree, and a grade B or better for master's degree and (if applicable) master's thesis (in the Norwegian grading system); b) documentation of English language proficiency. See OsloMet regulations for more detail.
What we offer
Excellent opportunities for performing high quality research as part of a highly competent and enthusiastic team of world-class researchers and engineers in an excellent innovative working environment. This also includes training to conduct independent research and publish in the top journals and conferences in the covered areas.
You will collaborate with top scientists in your field and have excellent prospects for personal development.
An informal and inclusive working environment.
Salary starting range from NOK 532.000
Generous support for travel and opportunities to build international networks, through established collaborations with industry, exchange programs and research visits with other universities, and funding to attend conferences.
Opportunities to build an international network through established collaborations with industry, exchange programs, and research visits to other universities.
Modern office facilities located in the heart of Oslo with an extensive public transport network within a few minutes’ walk and a vibrant city life right outside the office doorsteps, with welfare schemes, and a wide range of sports and multiple cultural offers.
Numerous other benefits, such as access to company cabin, BabyBonus arrangements, sponsored social events, generous equipment budgets (incl. laptop, phone and electronic communication subscriptions), comprehensive travel/health insurance policy, etc.
Relocation assistance: accommodation, visas, free Norwegian language courses for employees and their partners/spouses, etc.
Excellent administrative research support
Wellness and work-life balance: Our employees’ health and well-being are a priority. We encourage flexible work arrangements to help balance work and home life efficiently.
SimulaMet is an equal opportunity employer, and women are particularly encouraged to apply.
Interested applicants are requested to submit the following:
1) Cover letter (max 2 pages): outline your motivation for applying to this specific position, for pursuing high quality research in the topics associated to this position, your research interests, and how/why you are qualified for the position (relevant experience, qualifications, etc.).
2) Transcripts from bachelor’s and master’s degrees (in Norwegian or translated to English) and copy of Master Thesis (if applicable). Please, note that a description of the grading system at the university/country where you took your degree must be attached. This must be an official document with a stamp from your university. Foreign diplomas must be translated into English by the university that issued the diploma. If you have not completed your master's degree at the time of application, you must attach a preliminary transcript in English or a Scandinavian language from your university by the application deadline. It is desirable that you also attach an overview of subjects/exams you will complete during Fall 2023. Relevant applicants must also send an official confirmation from the educational institution by November 30, 2023 that all examinations for the master's degree, including the master's thesis, have been completed.
3) Scientific work (if applicable) that you want to be assessed.
5) A list of up to two references with contact information.
Official diplomas and transcripts must be submitted in any case before taking up the position and no later than December 1st, 2023. If your educational institution is not able to deliver an official diploma by this deadline, you must submit documentation from the institution that confirms that your master´s degree is completed by the same deadline.
Applications will be evaluated on a rolling basis, with a deadline of October 8, 2023. The PhD position will be filled as soon as possible.
Start Date: As soon as possible, and December 1st, 2023, at the latest, unless agreed otherwise.
According to the Norwegian Freedom and Information Act (Offentleglova) information about the applicant may be included in the public applicant list, also in cases where the applicant has requested non-disclosure.
Inquiries about the PhD position can be made by email to:
Chief Research Scientist/Research Professor Baltasar Beferull-Lozano (email@example.com),
Head of the SIGIPRO Department
If you have administrative questions about the position, please contact Research Advisor Jennifer Hazen (firstname.lastname@example.org).
If you would like to apply for the position, you must do so electronically through our recruitment system.