Accurate interpretation of the online learning model for 6G-enabled Internet of Things

The next-generation network (6G) has more strict requirements for the online learning ability and high interpretability of the learned systems. Machine learning is expected to be essential to assist in making the networks efficient and adaptable, but most promising methods often are treated as "black boxes" due to the deep structures and high nonlinearity. Therefore, this paper attempts to study the interpretations of machine learning algorithms to make them more applicable to the 6G-enabled internet of things (IoT) networks. Typically, this paper focuses on the new model GBDT2NN, which distills the knowledge learned by gradient boosting decision tree (GBDT) into neural network (NN) models to retain the learning ability of numerical data and rise the ability of online learning at the same time, but it loses the interpretability. This paper conducts an empirical study on explaining individual prediction of GBDT2NN by taking use of the feature importance learned from GBDT, and then further explores whether the explanation can improve the approximation process. In addition, this paper proposes two methods to obtain the interpretations: the Independent method and the Joint method. Experiments on several datasets of IoT networks show that the proposed methods can achieve better performance on both explanations and predictions.