In partnership with Intel® Software and Services, Center for Scientific Computing (CSC) of the São Paulo State University (UNESP) became in 2017 a Center of Excellence in Machine Learning (CoE-ML). These centers are part of the Intel® AI Academy, a global initiative that fosters collaboration between the company and universities with the aim of teaching, training and developing research and applications in Artificial Intelligence.
Within the scope of this project, CSC provides courses and workshops for the scientific community and software developers. In addition, Proofs of Concept (PoC) involving Machine Learning are conducted, such as classification of images using neural networks and performance analysis of tools optimized for an Intel® architecture. To request a workshop or a PoC, contact us!
A little bit about
Center for Scientific Computing (CSC)
The Center for Scientific Computing at UNESP was created on November 12, 2008 with the purpose of hosting important academic projects linked to the State University of São Paulo, such as GridUNESP and the São Paulo Regional Analysis Center (SPRACE).
The objectives of CSC include: expand and integrate the computing capacity of UNESP, develop partnerships with universities and private institutions, consolidate collaborations with international groups, develop human resources in high performance computing, and create conditions for the scientific and technological advancement of the country.
Science Is Not an Individual Sport
Machine Learning for Boosted Jet Classification in High Energy Physics
Jets are clusters of particles. Unsupervised learning Jet classification . Image recognition. Convolutional Nets.  J. R. Vargas, R. M. Cobe, S. F. Novaes, and T. R. Tomei, “Boosted jet identification with machine learning,” in National Meeting of Particle Physics, Passa Quatro-MG, Brazil, September 2017.
Identification of Cars (Brand and Model)
Objectives Explore possible neural architectures to handle the problem of car model identification. Reproduce the results reported in the paper linked here. Datasets The Comprehensive Cars (CompCars) dataset Stanford dataset Tools/Frameworks GoogLeNet Caffe Model. Tensorflow for poets. Expected Results Reproduce accuracy scores reported in the CompCars paper. Deliver…