Adrián Alcolea Moreno

I am a Ph.D. student at the Computer Architecture Group (gaZ) at the University of Zaragoza under the supervision of Javier Resano, where I am working on the development of hardware support for DNNs.

Before that, I graduated in Social Work from the University of Zaragoza, Spain (2010), I obtained my master degree in International Peace, Conflict and Development Studies from the University Jaume I of Castellón, Spain (2012), and then I graduated in Computer Engineering from the University of Zaragoza, Spain (2017).

Email  /  CV  /  Biography  /  LinkedIn

Current Research Topic

I am interested in the development of efficient hardware support for Neural Networks. In particular, Convolutional Neural Networks used for the classification of Hyperspectral images.


An Efficient Hardware Accelerator to Handle Compressed Filters and Avoid Useless Operations in CNNs
Adrián Alcolea, Javier Olivito, Javier Resano
ACACES poster abstract, 2018
bibtex  /  poster

Due to sparsity, a significant percentage of the operations carried out in Convolutional Neural Networks (CNNs) contains a zero in at least one of their operands. We propose the implementation in an FPGA of an architecture for CNNs capable of taking advantage of sparsity in two different ways: compressing sparse matrices and avoiding multiplications with zero in their operands.


On-board Efficient Hyperspectral Image Classification Using Convolutional Neural Networks on Reconfigurable Hardware
Adrián Alcolea, Javier Resano
ACACES poster abstract, 2017
bibtex  /  poster

The aim of this research is to develop efficient solutions capable of processing hyperspectral images on-board at real-time. For its achievement, we will explore the use of CNNs and their optimization trough the development of specific hardware support, based on FPGAs.

I like this website