With the rapid development of cyberspace and smart home technology, human life is changing to a new virtual dimension with several promises for improving its quality. Moreover, the heterogeneous, dynamic, and internet-connected nature of smart homes brings many privacy and security difficulties. Unauthorized access to the smart home system is one of the most harmful actions and can cause several trust problems and relationship conflicts between family members and invoke home privacy issues. Access control is one of the best solutions for handling this threat, and it has been used to protect smart homes and other Internet of Things domains for many years. This survey reviews existing access control schemes for smart homes, which concern the essential authorization requirements and challenges that need to be considered while designing an authorization framework for smart homes. Furthermore, we note the most critical challenges that other access control solutions neglect for smart homes.

The Extreme Learning Machine (ELM) is an effective learning algorithm for a Single-Layer Feedforward Network (SLFN). It performs well in managing some problems due to its fast learning speed. However, in practical applications, its performance might be affected by the noise in the training data. To tackle the noise issue, we propose a novel heterogeneous ensemble of ELMs in this article. Specifically, the correntropy is used to achieve insensitive performance to outliers, while implementing Negative Correlation Learning (NCL) to enhance diversity among the ensemble. The proposed Heterogeneous Ensemble of ELMs (HE ${}^{\mathrm{2}}$LM) for classification has different ELM algorithms including the Regularized ELM (RELM), the Kernel ELM (KELM), and the $L_{\mathrm{2}}$-norm-optimized ELM (ELML2). The ensemble is constructed by training a randomly selected ELM classifier on a subset of the training data selected through random resampling. Then, the class label of unseen data is predicted using a maximum weighted sum approach. After splitting the training data into subsets, the proposed HE ${}^{\mathrm{2}}$LM is tested through classification and regression tasks on real-world benchmark datasets and synthetic datasets. Hence, the simulation results show that compared with other algorithms, our proposed method can achieve higher prediction accuracy, better generalization, and less sensitivity to outliers.