Master thesis: Reinforcement learning for link adaptation
As a leader in wireless communications research and development, Ericsson is constantly exploring new approaches that fully utilize the available radio resources at the highest possible performance level. The challenge of continuing to provide best-in-class solutions for 5G and beyond radio networks will require handling a large number of diverse use cases such as extreme data rates for mobile broadband and ultra-low latencies for time-critical control applications. A promising new research direction in the optimization of future radio networks is data-driven machine learning that can autonomously tune the system parameters. These machine learning techniques are further motivated by the recent advances in this area, viz. deep learning, as well as the availability of large amounts of labeled and unlabeled data generated in cellular networks.
Adapting the radio parameters in real time to optimize the link performance, known as link adaptation, is a challenging task. Current link adaptation techniques rely on theoretical models and heuristics that are often sub-optimal in practical deployments. Recently, machine learning techniques that learn the link characteristics online have been proposed to improve link adaptation performance. In particular, reinforcement learning (RL) may be used to sequentially explore the radio parameter space and exploit this knowledge to select parameters that satisfy some optimality criteria related to the link error rate and throughput.
This thesis work will investigate RL techniques for online data-driven link adaptation. In this context, the thesis will propose and evaluate RL approaches for link adaptation in simulated radio environments and potentially within a testbed setup. The starting point for RL techniques will be multi-armed bandits that allow closed-form bounds on performance. Additionally, other promising RL and non-RL approaches may also be investigated. The substantial thesis goals will span the relative performance and robustness of evaluated RL techniques compared to traditional techniques for link adaptation.
The successful candidate must have:
· Strong understanding of wireless communication systems and techniques.
· Familiarity with machine learning in general and reinforcement learning in particular is a plus.
· Familiarity with machine learning tools such as k-NN, SVMs, neural networks, or random forests, etc., is a plus.
· Excellent programming skills in more than one general purpose or scripting language (e.g., Python, Java, C/C++).
· Familiarity with related supporting tools (e.g., Eclipse, Vim, Emacs), compiling environments in the major operating systems (e.g., gcc, Visual C++) or emulation techniques (virtual machines, Cygwin, MinGW).
· Excellent communications skills.
· Fluent in English, both written and spoken.
· Self-motivated and positive attitude.
This position is for one or two students. Scope is for 30 university credits (Swedish högskolepoäng).
Kista, Stockholm, Sweden
Preferred Starting Date
Please include your most recent grades in your application.
The selection process is ongoing, please apply as soon as possible. Last application date: 2018-11-23
For any further questions regarding this position, please contact:
Mentor: Vidit Saxena email@example.com
Recruiter:Linnéa Korneliusson firstname.lastname@example.org
Ericsson provides equal employment opportunities (EEO) to all employees and applicants for employment without regard to race, color, religion, sex, sexual orientation, marital status, pregnancy, parental status, national origin, ethnic background, age, disability, political opinion, social status, veteran status, union membership or genetics.
Ericsson complies with applicable country, state and all local laws governing nondiscrimination in employment in every location across the world in which the company has facilities. In addition, Ericsson supports the UN Guiding Principles for Business and Human Rights and the United Nations Global Compact.
This policy applies to all terms and conditions of employment, including recruiting, hiring, placement, promotion, termination, layoff, recall, transfer, leaves of absence, compensation, training and development.
Ericsson expressly prohibits any form of workplace harassment based on race, color, religion, sex, sexual orientation, marital status, pregnancy, parental status, national origin, ethnic background, age, disability, political opinion, social status, veteran status, union membership or genetic information.
Primary country and city: Sweden (SE) || || Stockholm || Stud&YP
Req ID: 263449