Post-Doctoral position : « Energy-efficient goal-oriented communication in networked control systems using machine learning tools »

Energy-efficient goal-oriented communication in networked control systems using machine learning tools

Date limite de candidature : 20/03/24
Date de début : 01/06/2024
Date de fin : 01/06/2025

Pôle : Télécoms et réseaux
Type de poste : Post-Doc ou ATER
Contact : Mohamad ASSAAD (

la fiche

Energy-efficient goal-oriented communication in networked control systems
using machine learning tools
In the last decade, communication systems have witnessed astronomical growth in both traffic
demand and widespread deployment. The interconnection and proliferation of devices with
their advanced sensing, computing, and learning capabilities are predicted to create an
astounding amount of data. A promising approach to reduce the amount of data transmitted over
future networks is to emigrate from the traditional communication paradigm in which
information relevant to semantic and communication goal is not considered in the
communications and network design. For instance, a typical autonomous vehicle collects from
its sensors up to several gigabytes of data per seconds. A machine-vision camera network is
usually composed of enormous number of cameras (e.g. millions of connected cameras in
metropolitan city) with large-size raw data generated by each camera. Clearly, only useful
information must be transmitted over communication networks and the usefulness is defined
with respect to the goal of the information. In this context, goal-oriented communication is
currently attracting interest from both academia and industry. One can define goal-oriented
communication as the framework dealing with how conveyed symbols affects the action of the
receiver in a desired way.
This postdoc opportunity lies at the intersection of communication, sensing and machine
learning in networked control systems. The system is composed of sources sending information
to a remote monitor. The focus is on investigating how a remote monitor/receiver can have
timely and accurate knowledge of the information required to achieve its goal. The monitor will
then decide when and how much information the source should send to achieve the goal while
satisfying a maximum energy constraint. The problem can be seen as an explorationexploitation
trade-off framework, for which, Restless bandit and reinforcement learning tools
could be useful to design an efficient solution.
Candidate: We are seeking a highly motivated candidate with solid mathematical background,
outstanding publication record and well-developed analytical skills. She/He is expected to
possess a solid background in communication theory, information theory and/or machine
learning. Knowledge in control theory will also be appreciated.
Supervisor: Mohamad Assaad
Laboratoire des Signaux et Systèmes, CNRS, CentraleSupélec, Université Paris-Saclay
Gif-sur-Yvette, France
[1] D. Gunduz, Z. Qin, I. E. Aguerri, H. S. Dhillon, Z. Yang, A. Yener, K. K. Wong, and C.-B.
Chae, “Beyond transmitting bits: Context, semantics, and task-oriented communications,”
IEEE Jorn. on Selected Areas in Comms., vol. 41, no. 1, pp. 5–41, 2023.
[2] Fan Zhou, Brahim Chaibdraa, Boyu Wang, “Multi-task Learning by Leveraging the
Semantic Information,” in AAAI conference, 2021.
[3] Saad Kriouile, Mohamad Assaad, “Minimizing the Age of Incorrect Information for
Unknown Markovian Source,”, journal paper under review,
[4] Maatouk, A.; Assaad, M.; Ephremides, A. The age of incorrect information: An enabler of
semantics-empowered communication, in IEEE Transactions on Wireless Communications,
[5] Hyowoon Seo, Jihong Park, Mehdi Bennis, Mérouane Debbah,”Semantics-Native
Communication with Contextual Reasoning,”
[6] Z. Weng, Z. Qin, X. Tao, C. Pan, G. Liu, and G. Y. Li, “Deep learning enabled semantic
communications with speech recognition and synthesis,” arXiv preprint arXiv:2205.04603,
[7] P. Young, A. Lai, M. Hodosh, and J. Hockenmaier. From image descriptions to visual
denotations: New similarity metrics for semantic inference over event descriptions. TACL,