Artificial Intelligence Resource Allocation with Safety Bounds in Wireless Networks
A postdoc position (18 months, extendable) is available in the IRISA laboratory at Rennes University (ADOPNET research team) on wireless communications, machine learning, radio resource scheduling and SDN. The research activities are supported by “SAFE” ANR project. Candidates with backgrounds in 5G, Artificial Intelligence, Scheduling and/or SDN are welcome.
The postdoc will collaborate with partners of the project such as researchers of IRISA, LabHC, XLim, Huawei, and QoS design.
When applied to communication networks, traditional approaches for control and decision-making require a comprehensive knowledge of system and user behaviors, which is unrealistic in practice when there is an increase in scale and complexity. Data-driven AI approaches do not have this drawback, but offer no safety bounds and are difficult to interpret. The SAFE project aims to design an innovative approach by combining the best of both worlds. In this new approach, intelligence is distributed in the network between a global AI (at the central level) and local AIs (at the edge level) collaborating with each other by integrating traditional models with graph neural networks and reinforcement learning. The approach, developed for partially or completely observable/controllable environments, will natively integrate safety bounds, interpretability and provide self-adaptive systems for routing, traffic engineering and scheduling.
Many scheduling or routing proposals as [Guerin21] and/or [Bordier21] provide efficient results in term of system capacity, fairness, energy consumption, etc. However, they are limited to system where knowledge of the system is highly significant. In order to provides solution for large scale in order to provide global optimization (vs. local), the postdoc research will focus on the design of new scheduling or routing strategies based on neural networks [Bethune20, Rusek20] or machine learning [Li18, Pham17, Pham19, Troia20] while offering safety bounds.
[Bethune20] L. Bethune, Y. Kaloga, P. Borgnat, A. Garivier, Amaury Habrard. “Hierarchical and Unsupervised Graph Representation Learning with Loukas’s Coarsening”. Algorithms, 13, 206. 2020.
[Bordier21] J-B. Bordier, C. Merlhe, P. Fabian, S. Baey, D. Garnaud, K. Bussereau, E. Livolant, C. Gueguen. “Buffer occupancy and link state opportunistic routing for wireless mesh networks”. Wireless Networks, 2021
[Guerin21] N. Guérin, M. Manini, R. Legouable and Cédric Gueguen, “High system capacity pre-scheduler for multi-cell wireless networks”. Wireless Networks, vol 27, pp 13–25, 2021.
[Li18] L. Chen, et al. “AuTO: Scaling Deep Reinforcement Learning for Datacenter-Scale Automatic Traffic Optimization”, In ACM SIGCOMM ’18, 2018, New York, USA, 191–205.
[Pham17] Pham Tran Anh Quang, K. Piamrat, Kamal Singh, C. Viho. “Video Streaming over Ad-hoc Networks: a QoE-based Optimal Routing Solution”. IEEE Transactions on Vehicular Technology, 66(2): 1533-1546, 2017.
[Pham19] Pham Tran Anh Quang, et al., “A Deep Reinforcement Learning Approach for VNF Forwarding Graph Embedding,” in IEEE Transactions on Network and Service Management, 16(4): 1318-1331, 2019.
[Rusek20] K. Rusek et al., “RouteNet: Leveraging Graph Neural Networks for network modeling and optimization in SDN”, IEEE Journal on Selected Areas in Communications, 38(10), 2260-2270, 2020. [Troia20] S. Troia et al., “On Deep Reinforcement Learning for Traffic Engineering in SD-WAN.” IEEE Journal on Selected Areas in Communications (2020).