Date: 1 July 2022
What will our industry look like in 10 years and what will we be researching?
Targeted treatment options will improve
Supercomputing Technologies Research Group - SC@FIT
I'm a bit of a sceptic, so I think there will still be a greater demand for computing power than we can meet. Anyway, there will be a gradual transfer of computationally intensive applications to supercomputers and the cloud, and people will use simpler and simpler terminals as their personal devices. Great advances will be made in medicine, particularly in the field of automated screening and diagnosis of various diseases, as well as the possibilities of precisely targeted treatment, which will be increasingly supported by software simulations and artificial intelligence.
Algorithms will better understand people
Speech Data Mining Research Group - SPEECH@FIT
"Easy" tasks, such as operating voice assistants in quiet conditions and with a restricted vocabulary in one language, or speaker recognition from long clean recordings, have already been solved, algorithms are commonly marketed and achieve "super-human" results. But in difficult conditions - noise, new languages, new domains or even dialects - humans still greatly outperform machines. In our field, we work hard to match human performance even under these difficult conditions. Much of this is aided by the "convergence" of machine learning - whereas a decade ago, speech research, computational linguistics, robotics, and even computer linguistics were separate fields, artificial neural networks have brought these fields together and developments in all areas have accelerated. In the future, we will definitely see more teacher-less training, automatic data crawling, multimodal approaches, and human user-centred development.
Crime will move even further into cyberspace
Accelerated Network Technologies Research Group - ANT@FIT
In 10 years, programmable networks will be a commonplace commodity. P4 programmability will become a common feature of switches and network cards. Network deployment at 1 Tb/s will be the order of the day. FPGA technology will make its way into processors as a means of accelerating applications similar to GPUs. Crime will move even further into cyberspace. Network security as well as network attacks will actively use machine learning and artificial intelligence. Based on specific cases of data misuse, people will be shocked at what can be learned and exploited from the available digital traces and social media accounts. There will be an increased push for privacy protection.
Mathematics must remain the fundamental lingua
franca of science
Formal Model Research Group - FM@FIT
We strongly believe that mathematics must remain the basic "lingua franca" of science, including computer science if it is to face the crisis of European science, which Edmund Husserl, a native of Prostějov and founder of phenomenology, warned about. The language of mathematics will remain thelanguage common to all of science. Therefore, we plan to continue to present our ideas in a strictly mathematical way; however, our focus will become more application-oriented. The main topic will be the computer processing of language, especially its analysis and translation. However, this orientation will no longer focus only on artificially created languages, such as programming languages, as has been the case so far. It will also focus on natural language translation, including the creation of a high-quality Japanese-Czech machine translator.
High Performance Computing Research Group - HPC@FIT
The Navier-Stokes equations are still among the seven unsolved problems of the millennium. Moreover, we deal with extending this problem to include more complex boundary conditions, which introduce considerable nonlinearity into the problems - making them even more difficult to solve numerically. Practically speaking, we would like to focus mainly on modelling the blood circulation in the human body. The idea is that, for example, a doctor would mark the blood vessel he wants to sever on a tablet during surgery, and a supercomputer would calculate the pressure distribution and flow after the procedure.
Autonomous vehicle navigation, augmented
reality and computational photography
Computational Photography Research Group - CPHOTO@FIT
Our methods will assist in the navigation of autonomous vehicles and drones. One of our applications - augmented reality projected directly onto the retina of the eye - will enable better orientation and understanding of nature for all age groups, including people with various disabilities. The photographic devices we have helped develop will capture a faithful "impression" of reality, which can also be freely altered without compromising credibility.
Network security based on the use of autonomous agents powered by artificial intelligence
Networks and Distributed Systems Research Group - NES@FIT
In the area of networking, virtualisation will be a major focus of attention, and the concept of programmable network devices will once again come to the fore. Here, we'd like to ride the wave of P4, which is a technology that big and important companies are pursuing. Another interesting topic is autonomous network security based on the use of autonomous agents powered by artificial intelligence.
Robots will collaborate with humans
ROBO@FIT Research Group
In the area of human-machine/robot communication and collaboration, it will be normal that robots will understand the production processes, equipment and objects in their environment, as well as the actions required of them. I do not see the future in complete autonomy of production robots, vehicles, drones, etc., but in the co-operation of these machines with humans, where both robot and human will do what each of them can do better and with less effort: the robot can perform repetitive actions, search for information and evaluate the situation, while the human can then define goals, possible solutions and make decisions. To do this, however, we need to find ways of effective communication between humans and machines. Even an inexperienced programmer can handle instructing a group of robotic manipulators and mobile robots in a natural way. Today we are trying to move towards this goal using, for example, augmented reality and automated situational awareness. We will see what possibilities will be brought about by technologies such as knowledge retrieval from the Internet, machine learning or reading brain activity.