abstract-In recent years, the streaming of AR/VR videos has seen a rapid increase due to its industrial and commercial applications in different fields like healthcare, education, fashion, etc. However, the streaming of AR/VR videos requires high Quality of Service (QoS) to provide adequate Quality of Experience (QoE) for individual users. AR/VR applications require stable and reliable connections with very low latency and high data rates to maintain a high user experience; end-user mobility can make the provision of these connections challenging. The integration of 5G and beyond with Network Slicing (NS) and Network Slice Management (NSM) enhances latency and throughput for critical applications by supporting URLLC (Ultra-low latency communication) and eMBB (Enhanced Mobile Broadband) service types [1] [2]. However, recent work in NS and NSM only considered spectrum sharing and management for different slice types but did not focus on the QoE of an individual user. This paper proposes a novel approach for Network Slice Management and Monitoring ($N S M^{2}$) for AR/VR streaming using Machine Learning (ML), which focuses on increasing the QoE and QoS for each user. In order to assess $N S M^{2}$ benefit, OpenAirInterface (OAI) is used for simulations. We generate a realistic dataset for evaluating and comparing non-ML-based approaches with our ML-based approaches. ML algorithms are evaluated for accuracy, recall, precision, and f-measure. Simulation results show that $N S M^{2}$ with the Convolutional Neural Network (CNN) model outperformed other solutions and achieved higher throughput and lower latency as well as improved QoE.