Orange Labs Networks, partner of SliceNet, has made a number of contributions in the area of Machine Learning and cognitive management to the ITU-T Focus Group on Machine Learning for Future Networks including 5G. These contributions included anomaly prediction and integration for an eHealth Use Case based on vertical feedback; and Noisy neighbour detection and integration in a virtualized infrastructure. Orange Labs Networks presented their  contributions and an overview of SliceNet architecture with focus on cognitive closed loop to the ITU-T Focus Group ML5G under study group 13 at theweb-conference, on 20th November 2019. See

The SliceNet Cognition Plane design and methodology is aligned with several of the main standardization bodies and associated documents with regard to Machine Learning and cognitive management.

Contribution (ML5G-I-198) proposes an anomaly detection model that serves as an intelligent QoE sensor by analyzing the data samples. The SliceNet Project aims to provide a network service for the vertical (national ambulance service) offering the best Quality of Experience (QoE) to paramedic teams. A smart connected ambulance is roaming through the network while sending data streams that should be delivered in real time with no quality degradation. A feedback mechanism is implemented, allowing for verticals to express the perceived quality of the service every second.

The anomaly detection model aims to observe the last 5 minutes of the perceived QoS in order to predict if a degradation in the network signal strength in the RAN segment may be perceived by the vertical in the following 5 minutes. Once signal quality is predicted to be degraded, remedial actions need to take place to maintain optimal QoE levels. By forecasting the signal strength degradation, network quality maintainance can be alerted in advance and the issue could be solved before it occurs. This will allow the perceived quality of the slices’ servicesto be maintained and it can allow the vertical to communicate feedback and supervise its “slice”.

SliceNet presents contributions at the ITU Focus Group Machine Learning 5G on anomaly prediction