The (HEA/HEAN)-n multilayer coatings consisting of alternating (AlCrFeMoTi)N (top) and AlCrFeMoTi layers were specially designed and successfully deposited on F/M steel substrate by reactive magnetron sputtering techn...
详细信息
Neural architecture search (NAS) has emerged as a transformative approach for automating the design of neural networks, demonstrating exceptional performance across a variety of tasks. Numerous NAS methods aim to opti...
详细信息
ISBN:
(数字)9798350359312
ISBN:
(纸本)9798350359329
Neural architecture search (NAS) has emerged as a transformative approach for automating the design of neural networks, demonstrating exceptional performance across a variety of tasks. Numerous NAS methods aim to optimize neural architectures within discrete or continuous search spaces, but each method possesses its own inherent limitations. Additionally, the search efficiency is notably impeded by suboptimal encoding methods, presenting an ongoing challenge. In response to these obstacles, this paper introduces a novel approach, evolutionary neural architecture optimization (ENAO), which optimizes architectures in an approximate continuous search space. ENAO begins with training a deep generative model to embed discrete architectures into a condensed latent space, leveraging unsupervised representation learning. Subsequently, evolutionary algorithm is employed to refine neural architectures within this approximate continuous latent space. Empirical comparisons against several NAS benchmarks underscore the effectiveness of the ENAO method. Thanks to its foundation in deep unsupervised representation learning, ENAO demonstrates a distinguished ability to identify high-quality architectures with fewer evaluations and achieve state-of-the-art result in NAS-Bench-201 dataset. Overall, the ENAO method is a promising approach for optimizing neural network architectures in an approximate continuous search space with evolutionary algorithms and may be a useful tool for researchers and practitioners in the field of NAS.
The effect of the Si addition content ranging from 3.3 to 9.1 at.% on the microstructure, mechanical properties and LBE corrosion behaviour of AlCrFeMoTi HEA coatings was investigated. The AlCrFeMoTiSix coatings still...
详细信息
Aluminum-silicon alloys are extensively utilized in aerospace and automotive industries owing to their exceptional wear and corrosion resistance. However, traditional casting and single wire arc additive manufacturing...
详细信息
When developing a blink input interface, conscious (voluntary) and natural (involuntary) blink types must be automatically classified. We previously proposed a method for blink type classification using a 3D convoluti...
详细信息
Helical cruciform fuels are novel in nuclear reactors, potential to increase reactor’s power density. However, the geometry is complicated so influence of it on the fuel performance is not identified yet. In this pap...
详细信息
The sensor signals collected by the nuclear power plant system equipment are mixed with complex noise information. The uncertainty of sensing data directly affects the application effect of artificial intelligence on ...
详细信息
It has been acknowledged that dominance resistant solutions (DRSs) often exist in the feasible region of multi-objective optimization problems. DRSs can severely degrade the performance of many multi-objective evoluti...
详细信息
Recently, a technology called floating nuclear power plant (FNPP) has been developed around the world to maximize the use of ocean and nuclear energy. The FNPP is required to investigate the impact of a ship collision...
详细信息
This paper aims to examine the applicability of "Extra drainage systems" which can complement conventional drainage system. Firstly, the authors asked participants about the difficulty using conventional dra...
详细信息
暂无评论