My main research area is the acceleration of
scientific computation through High Performance Computing,
mainly using GPU computing. I am interested in
applying and developing parallel solutions in several fields,
such as the simulation
of bioinspired computing models (Membrane
Computing and Multi-Agent Systems),
training and deployment of Machine Learning models (Deep
Learning in image and video understanding), and image
coding algorithms (JPEG2000
for DCP and IMF, and JPEGXS for
lightweight and low latency compression).
So far I have been working on the acceleration
of applications by leveraging GPUs with CUDA. At the
Research Group on Natural Computing (University of Seville), I
have developed the first simulators for P systems on CUDA. At
the Department of Moving Picture Technologies (Fraunhofer
IIS), I contributed to the acceleration of JPEG2000 codecs
with CUDA and to the standardization of JPEGXS for its
parallelization on GPUs. Furthermore, I am working with Deep
Learning and its potential applications to image and
video understanding, automatic sign language translation and
production, pose estimation and signal processing. I am also
studying the parallelization of new Multi-agent Systems
(e.g. for Crowd Simulation). I visited Matthias
Niessner's group at TUM in order to study how to accelerate
sparse convolutions on GPUs.
I am also interested in researching new design
patterns for parallel scientific computations, in order
to help adapting and implementing them to GPU architectures.
Other topics I would like to study are also related to the
fields of Data Science, High Energy Physics and
Astrophysics. Although it is a diverse set of interests,
I just love science and its applications.
In summary, my research topics today are:
- GPU Computing (CUDA and standards such as OpenCL, Vulkan
and SYCL)
- Parallel Computing (High Performance Computing)
- Deep Learning Applications (Image and Video Understanding)
- Machine Learning Acceleration (sparse convolutions)
- Simulation and Modeling of Complex Systems (Population
Dynamics, Multi-agent systems)
- Bio-inspired Computing (Membrane Computing)
- Acceleration of Image Coding (JPEG200, JPEGXS)
