PhD "AI for Semantic Communications in 6G" F/M
ref :2023-24764 | 05 Sep 2023
apply before : 02 Oct 2023
28 chemin du Vieux Chene 38240 MEYLAN - France
about the role
Your role is to perform a PhD thesis work on "AI for Semantic Communications in 6G".
In the context of future 6G networks, initial work announces even higher requirements in terms of throughput, latency and capacity than for 5G. At the same time, new societal objectives are being put forward, such as low-cost global coverage, privacy and a smaller energy footprint [1].
In this context, and with the objective of improving the efficiency of our transmissions, the subject of semantic communications is developing [2]. The principle is to take into account the nature of the data and to transmit only the information that is relevant to the purpose of the communication. This approach is a paradigm shift from current communications, which measure their performance by bit or packet error rate metrics, and not by the achievement of the transmission goal. By choosing this goal for its optimization, semantic communications promise both better compression of useful information, but also better resistance to errors. Thus, current approaches use AI techniques (deep learning) to define the semantic projection space of the data: the embeddings. The semantic space must therefore be optimized to both compress the initial information, while being robust to the disturbances of the radio propagation channel.
The objective of the thesis is to study the automatic construction and the transmission methods of embeddings for the realization of a communication objective.
The main technical issues to be addressed are:
- The automatic construction of the latent space. This implies the extraction of salient features and a relevant compression of the data to fulfill the communication objective, but also a resistance to transmission errors and a consideration of the propagation channel.
- The different transmission methods for embeddings: from discrete constellations to continuous approaches.
- The definition of metrics to evaluate the quality of the semantic space construction and the degree of realization of the communication objective.
- The comparison with standard transmission methods, traditionally not evaluated on the communication goal.
The thesis will focus on different types of multimedia content, progressively increasing the complexity of use cases, from text transmission to AR applications.
A progressive abstraction on the nature of the data and on the final goal of the communication will be studied in the context of privacy.
about you
You are a graduate or future graduate of an engineering school and/or a research master (BAC+5), with a specialization in data science and/or telecommunications.
You have the following skills:
- You have a very good knowledge of neural networks, and more generally of machine learning techniques.
- You have a solid knowledge of radio signal processing.
- You have significant experience in developing in Python, especially with Pytorch and/or Tensorflow libraries.
- You are fluent in English. The work done in the framework of this thesis will be the subject of patent applications and publications in international conferences.
- You are methodical, curious, autonomous and motivated to tackle an open research topic.
- You have good interpersonal skills and know how to communicate on your subjects of interest.
A first experience in the world of research is a plus (R&D internship, scientific paper writing).
The position is expected to start in October - November 2023.
additional information
Semantic communications launch a new field of study and promise to change the way we conceive our communication systems in numerous cases.
Although the construction of the latent space for a given task is a common operation in the field of machine learning, the addition of a communication system at the center challenges the usual construction mechanics of this projection space. Limited energy quantity, quantization and propagation channel are all constraints that impact the way to design an efficient latent space. Moreover, the optimization of the transmission of the embeddings is also at stake, including an end-to-end optimization approach, and opening the perspective of more energy efficient transmissions.
This thesis is supervised by Louis-Adrien Dufrène, Grégoire Lefebvre and Quentin Lampin at Orange Innovation in Meylan (38), having supervised the thesis "AI models for digital signal processing in future 6G-IoT networks" of G. Larue.
As the thesis is co-supervised by an industrialist and an academic, you will be required to conduct your research in both environments, thus giving a variety of career options.
Bibliography
[1] "Orange's vision for 6G", White Paper, March 2022.
[2] Gündüz, Deniz, et al. "Beyond transmitting bits: Context, semantics, and task-oriented communications." IEEE Journal on Selected Areas in Communications, 2022.
department
Orange Innovation brings together the research and innovation activities and expertise of the Group's entities and countries. We work every day to ensure that Orange is recognized as an innovative operator by its customers and we create value for the Group and the Brand in each of our projects. With 720 researchers, thousands of marketers, developers, designers and data analysts, it is the expertise of our 6,000 employees that fuels this ambition every day.
Orange Innovation anticipates technological breakthroughs and supports the Group's countries and entities in making the best technological choices to meet the needs of our consumer and business customers.
Within Innovation, you will be integrated in a research team with a dual expertise in artificial intelligence and radio connectivity. You will be surrounded by researchers working on innovative and fundamental subjects for the future of telecommunications and in particular for 6G. The team also has a radio laboratory and expertise to perform field experiments.
contract
Thesis
Only your skills matter
Regardless of your age, gender, origin, religion, sexual orientation, neuroatypia, disability or appearance, we encourage diversity within our teams because it is a strength for the collective and a vector of innovation. Orange Group is a disabled-friendly company: don't hesitate to tell us about your specific needs.
Similar offers
Orange SA
Orange Group
of our employees are proud to work for Orange
recommend Orange as a good place to work
is the candidate experience in France, in the category of companies with over 1,000 employees
Since 2011, Orange has GEEIS (Gender Equality European & International Standard) certification in some twenty countries