6G Scenario-Aware and DRL-Enabled Approach for Explainable Recommendation

6G heralds a paradigm shift from network performance to operational mechanisms, with on-demand services standing as its quintessential feature. Recommendation systems, serving as sophisticated tools for knowledge analysis and correlation, efficaciously address the dynamic exigencies of service demands. However, the conventional template-based scenario setting in mobile communication networks proves inadequate for the multifaceted and correlated requirements inherent to 6G networks, thereby imposing elevated standards on the precision and adaptability of scenario cognition. Concurrently, explainability emerges as a salient characteristic of 6G networks. Confronting the opacity of vast services and artificial intelligence algorithms, the challenge lies in discerning reliable on-demand strategies. To tackle the aforementioned issues, we propound an explainable recommendation framework grounded in 6G scenario awareness. Subsequently, we operationalize 6G feature representation based on the Scenario-aware Mutual Learning (SAML) model, intricately modeling the interconnectivity between scenarios. Lastly, we employ reinforcement learning methods to achieve explainable recommendations. Rigorous experiments on our constructed 6G service dataset attests to the superior performance of our recommender system compared to baseline methodologies.