On Redundancy and Diversity in Cell-based Neural Architecture Search

Abstract

Searching for the architecture cells is a dominant paradigm in NAS. However, little attention has been devoted to the analysis of the cell-based search spaces even though it is highly important for the continual development of NAS. In this work, we conduct an empirical post-hoc analysis of architectures from the popular cellbased search spaces and find that the existing search spaces contain a high degree of redundancy: the architecture performance is minimally sensitive to changes at large parts of the cells, and universally adopted designs, like the explicit search for a reduction cell, significantly increase the complexities but have very limited impact on the performance. Across architectures found by a diverse set of search strategies, we consistently find that the parts of the cells that do matter for architecture performance often follow similar and simple patterns. By explicitly constraining cells to include these patterns, randomly sampled architectures can match or even outperform the state of the art. These findings cast doubts into our ability to discover truly novel architectures in the existing cell-based search spaces, and inspire our suggestions for improvement to guide future NAS research. Code is available at https: //github.com/xingchenwan/cell-based-NAS-analysis.

A Comprehensive Survey of Cybersecurity Threats, Attacks, and Effective Countermeasures in Industrial Internet of Things

“…For older, un-upgraded systems in the high-complexity environment under study, the complicated relationship descriptions and the minute distinctions that set them apart exhibit multivariate standardization systems while maintaining high heterogeneity [7]. The intricacy and secrecy of supply chain attacks, which make them difficult to combat effectively, have made them stand out among the other dangers as a considerable worry [131,132]. Approaches to risk management are frequently used to lessen such attacks.…”

“…One of the advanced security objectives of information is cryptography, which has subobjectives including confidentiality, non-repudiation, integrity, and authenticity. In manuscript [132], the authors proposed a highly efficient quantum digital signature (QDS) protocol that uses asymmetric quantum keys. The authors successfully completed the demonstration of the complete function of a quantum security network, which met all required objectives of information security.…”