HPC is utilized in our work on the reconstruction of images from neural network weights in the domain of adversarial machine learning. The data reconstruction attack involves images with many pixels and a significant number of iterations for the attack. This requires intensive computation. Using HPC, we reduced the processing time from 2.5 days to 4-6 hours, depending on the experiment.
Big data services are an essential part of our educational curriculum, supporting students in their research and courses on information retrieval and the big data course itself. Students learn to process tera/petabytes of data using big data technologies and apply them to real-world applications like intrusion detection, web search index preparation, and graph theory.
Cloud VMs are extensively used by students and faculties for various research and experimentation projects. The applications of cloud VMs range from different programming languages, information security experiments to computation-intensive tasks like deep learning. The cloud facility supports students who have low-performance machines, enabling them to carry out their tasks smoothly.
|
|
|
|