Data Sciences and Artificial Intelligence
The DSAI research team was created with the start of the research department at UFAZ on 01 Sep 2021. The DSAI research team is mainly working on Massively Parallel Evolutionary Optimization problems. As well as the analysis of Big Data complex networks.
Massively Parallel Evolutionary Optimization
In the last few years, there has been a massive increase in the size and complexity of data sets in many different scientific disciplines as a result of advances in technology. Large and complex data sets are usually found in fields such as medical imaging, physics, material science, remote sensing, etc. Working with such data sets has become very challenging since traditional/standard methods are no longer efficient. As a result, new techniques must be used to analyze large and complex data. In our team, we use a new harmonic analysis approach that uses artificial evolution, and experiment it on large noisy simulated and real data coming from Fourier Transform Ion Cyclotron Resonance (FT-ICR) mass Spectrometry. The evolutionary method and algorithm we propose and work on (sinus-it) offers both a better tolerance to noise and the possibility to directly find the phase of the different sines, which (in the case of FT-ICR) can be used to help find the finer peaks of the compound fine structure analysis.
Big Data complex network analysis
Our societies can be described as a network of people connected by friendship, familial or professional relations. The Internet is a network of routers connected by physical data connections. Twitter is a network of 211 million daily active users (Q3 2021) linked together by friendships and retweets. About 10e11 neurons are connected by biochemical reactions in human’s brain. Wikipedia is a network of millions of web pages linked together by hyperlinks (>54M pages in English Wikipedia by Dec 2021). Protein interactions, network of Hollywood actors, power grids, highways and many other fundamentally different systems existing in real life have a network architecture. The spread ubiquity of networks makes their study indispensable. In our research we mainly focus on using the well known Google matrix (GM) and Reduced Google matrix (RGM) that we are one of the pioneers worldwide to offer and work on. We use those methods in order to calculate the influence of nodes as well as the hidden links that could be investigated showing the impact of connectivity on some specific clusters and the whole complex network.
Moreover, a new machine learning research work is being established on X-RAY dental images. The ongoing task is to analyse the differential diagnosis of jaw cysts using machine learning on panoramic X-RAY.
Experimentations can be run on our facilities, composed of state-of-the-art hardware
our specific networking lab, equipped with professional hardware for hands-on practice,
individual Intel Core i9-7900X ten core 64GB computers equipped with 2x ZOTAC GeForce RTX 2080 Ti GPGPU cards,
a non-newtonian supercomputer made of the same machines, working together as a group, creating a > 800 TFlops PARSEC machine for compute-intensive research projects
Yunuslu Ibrahim (M2- Intern)
Alizada Aliisa (M2- Intern)
Yusibov Sanan (M2- Intern)
Jafarli Kanan (M2- Intern)
The DSAI team offers research internships for second year master degree students (5 months) as well as last year bachelor degree students (3 months). Internship topics are purely devoted to research and could be performed either at UFAZ or with collaborators. After doing their internships, students will be ready to apply to different PhD programs worldwide.