PUMPS Summer School, July 11-15, 2016

NVIDIA kindly donated 2 K80 GPUs for the best poster and best clinic awards.

Best Poster Awards

We would like to congratulate the winner of the 2016 Best Poster Award

Title: Chunking SVM Training Implementation in CUDA
Author(s): Josef Michálek, Jan Vaněk
Affiliation: University of West Bohemia
Poster: Chunking SVM Training Implementation in CUDA

Support Vector Machines are one of the most popular tools for classification or regression. There are several open-source SVM training implementations utilizing GPUs, but many of them are now obsolete and don't perform well on modern GPUs. There haven't been much development in this area in the past few years and the goal of this work is to introduce our own implementation of SVM training in CUDA. This poster shows the basic workings of our implementation and compares the training speed and model quality to several well-known open-source SVM training implementations.

And the runner-up and Best Poster Finalist

Title: Stereo Matching using SGM on the GPU
Author(s): Daniel Hernandez-Juarez, Alejandro Chacon, Antonio Espinosa, David Vazquez, Juan Carlos Moure, Antonio M. Lopez
Affiliation: Universitat Autònoma de Barcelona
Poster: Stereo Matching using SGM on the GPU

Dense, robust and real-time computation of depth information from stereo-camera systems is a computationally demanding requirement for robotics, advanced driver assistance systems (ADAS) and autonomous vehicles. Semi-Global Matching (SGM) is a widely used algorithm that propagates consistency constraints along several paths across the image. This work presents a real-time system producing reliable disparity estimation results on the new embedded energy efficient GPU devices. Our design runs on a Tegra X1 at 42 frames per second (fps) for an image size of 640×480, 128 disparity levels, and using 4 path directions for the SGM method.

Best Clinic Award

Title: CUDA Accelerated Wavelet Neural Network for Time Series Forecasting
Author: Athanasios M. Kintsakis
Affiliation: Aristotle University of Thessaloniki, Greece
Looking to improve the Particle Swarm Optimization training process of a Wavelet Neural Network Athanasios created in his master thesis by introducing randomness in order to avoid premature convergence in local optima.

awards.txt · Last modified: 2016/07/28 12:25 by apena
Drupal Garland Theme for Dokuwiki