Recently neural architecture search (NAS) has raised great interest in both academia and industry. However, it remains challenging because of its huge non-continuous space. Instead applying evolutionary algorithm or reinforcement learning as previous works, this paper proposes a direct sparse optimization NAS (DSO-NAS) method. The motivation behind DSO-NAS is to address the task view model prun...

We study the problem of computing the k-visibility region in the memory-constrained model. In this model, the input resides in a randomly accessible read-only memory of O(n) words, with O(logn) bits each. An algorithm can read and write O(s) additional words of workspace during its execution, and it writes its output to write-only memory. In a given polygon P and for a given point q ∈ P , we sa...

Writing, as a productive skill, requires an accurate in-depth knowledge of the grammar system, language form and sentence structure. The emphasis on accuracy is justified in the sense that it can lead to the production of structurally correct instances of second language, and to prevent inaccuracy that may result in the production of structurally erroneous sentences. The present research intend...

Aberration correction in the TEM is now relatively well established with more than 20 corrected instruments installed worldwide. This paper will review these instrumental advances and will also describe the underlying theory and computation required to optimise data acquisition and interpret aberration corrected images. Finally the combination of direct electron optical correction and indirect ...