Accelerating neural network architecture search using multi-GPU high-performance computing

Lupión, Marcos; Cruz, Nicolás C.; Sanjuan, Juan F.; Paechter, Ben; Ortigosa, Pilar M.

doi:10.1007/s11227-022-04960-z

Cited by 4 publications

(2 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Among them, nature-inspired meta-heuristics, such as genetic algorithms, are widespread [12][13][14]. However, the intricate set of constraints makes it difficult to apply this kind of method naturally [17,35,38]. A plain black-box box-constrained problem is significantly more convenient., i.e., a formulation defining bound constraints only [11,15].…”

Section: Problem Statementmentioning

confidence: 99%

“…Nonetheless, it is relevant to mention that there is a price to pay: Different yet close values of components c 5i−4 result in choosing the same region. This situation might cause plateaus in the objective function, i.e., different points with the same value, as it may occur with the encoding method used in [38] for describing neural network architectures. Nonetheless, as the methods considered are derivative-free, their exploratory strategy is not affected by this situation, and the benefits exceed the drawbacks.…”

Section: Details On the Auxiliary Objective Functionmentioning

confidence: 99%

See 1 more Smart Citation

Automatic placement of cameras in a particle accelerator using a 3D engine and computational optimization

Cruz,

Rouret,

Ortigosa

et al. 2024

Preprint

View full text Add to dashboard Cite

Particle accelerators have high-radiation parts requiring remote maintenance tasks, which involve using cameras generally placed by human experts. This work addresses camera placement through black-box global optimization. This approach allows opting for automated solutions and relieving experts of that responsibility. The formulation of the objective function hides the environment-specific placement constraints and shows domain ones only. The proposed methodology allows the use of regular meta-heuristics for standard black-box box-constrained problems, as it is only necessary to honor the bounds of variables. Accordingly, the optimization algorithm used is the genetic algorithm provided by the Global Optimization Toolbox of MATLAB. The objective function computation relies on the simulation capabilities of the widespread Unity game engine. It brings us an integrated framework for defining the target process to consider and for checking the different proposals visually. Based on this framework, this work reconstructs the virtual reality simulation of a maintenance task in a particle accelerator. The results obtained show that the camera positioning achieved by the proposal for four cameras and one target process outperforms the arrangement defined by a human expert.

show abstract