Artificial intelligence does not necessarily refer to a robot. It can simply be a piece of computer software. An algorithm that can make decisions on its own. Artificial intelligence methods are already actively being applied by Lithuanian scientists in various practical fields – from automatic surface defect detection and early detection of machine-tool malfunctions to medicine and defence projects.
In Imitation of Human Ability
“The inquiry we get the most is about product surface defects. You can get these defects on various surfaces – furniture boards, linen fibre or any kind of textile. We develop systems that allow the automatic detection of such defects using cameras and, of course, algorithms,” explains Vidas Raudonis, PhD of the Faculty of Electricity and Electronics at Kaunas University of Technology (KTU).
Developing an algorithm that can visually identify a defect is difficult work. Imitation of the human ability of seeing otherness in a uniform environment is possible due to artificial neural networks. For about six years now, KTU researchers have been developing visual quality inspection systems. Vakarų Medienos Grupė is one of the largest local companies the KTU team works with.
It also successfully collaborates with international companies such as Siemens, Klinkmann and Schneider Electric. Their clients are trained in industrial controller programming courses. The software they install controls certain industrial processes, for example, the production of curd cheeses or beer. According to Vidas Raudonis, production would now be difficult to imagine without such controllers. They have become the basis of the fourth industrial revolution. Through these controllers, information travels from the lowest levels of production to the highest levels of production planning and resource distribution. The flow of data inside the controller can be likened to an artificial intelligence that can make decisions when it needs to filter essential information and model different situations.
Vitally Important Helpers
Several years ago, KTU scientists founded Power of Eye, a company dedicated to helping people with disabilities. The eye movement tracking system they created allows a person who has no control of their arms or legs to use their eye movements to control a computer and a device connected to the computer, type text and perform simple acts of communication if the person cannot speak.
Along with AT Medical, KTU researchers set up an incubator for growing human embryos up to the 8-16 cell stage. In fertility clinics, artificially fertilised eggs are kept in an incubator for up to five days until they develop into embryos. The new device captures images of the embryos every five minutes, helping specialists keep track of their development. The embryologist no longer needs to take the embryos out of their warm, controlled environment into a colder room-temperature environment in order to assess them under the lens of a microscope.
“Artificial intelligence assists the embryologist by letting them assess the embryos’ vital functions in a non-stressful environment. Having analysed up to 100 thousand images, it takes responsibility for marking those critical areas, so that the embryologist pays attention to them, and no detail goes unnoticed,” explains Raudonis, speaking of an artificially intelligent aid that has already entered the international market.
More Accurate, More Insightful
According to Professor Tomas Krilavičius, head of the Applied Informatics Department of the Faculty of Informatics at Vytautas Magnus University (VDU), VDU specialists are conducting research in cooperation with businesses from the engineering industry – Baltec CNC and Aedilis – to develop technology that can detect machine-tool malfunctions early on and prevent various technical issues related to their use. Process history is accumulated by observing how the systems work, allowing manufacturers to replace the old equipment before it malfunctions.
Several years ago, a unique medical device was developed in collaboration with the advanced technology company Rubedo Sistemos. The device was designed to compensate for the respiration-induced movement of cancer patients during radiotherapy. As the patient breathes, the tumour moves as well. The new device monitors how the patient breathes and directs the radiation with sub-millimetre accuracy exactly where it needs to go, without damaging healthy tissue and reaching previously inaccessible areas.
A partnership with Terra IT, an IT company, resulted in successful research on how to help stores order stock. The technology records the store’s sales history: when what products are bought and how many products are bought, depending on weather conditions, time of year and store location. On the basis of this data, methods are developed that allow store managers to better estimate what products to order and in what quantities, so that their store is sufficiently stocked but not over-stocked.
“Another relevant field of artificial intelligence application is propaganda and defence”, notes Professor Krilavičius, “For example, for monitoring various information sources, so that you could identify who is spreading propaganda and how, what methods they are using to indirectly discredit others and create a bad background around them.”
Such research is being conducted by specialists from the VDU Faculty of Informatics, the Baltic Institute of Advanced Technologies and other institutions in cooperation with NATO partners as part of an initiative launched by the NATO Science and Technology Organization.
What Man Can’t Do
Vilnius Gediminas Technical University (VGTU) and a Lithuanian transportation company is getting ready to submit one of its largest projects to the Intelektas LT-2programme. The project is dedicated to railway logistics management, monitoring and prevention, and is also closely connected to the internet of things. As Professor Antanas Čenys, the prorector for Research and Innovation at VGTU, explains, many sensors monitor the railway and locomotives, generating data. A human being would not have the capacity to process this data, which is why computer systems that allow fast decision-making could help them make these decisions.
A similar project is being prepared for submission with a leading European and the largest Lithuanian transport and logistics company. Human personnel can no longer keep up and manage the massive flows of cargo vehicles. Decisions need to be made by artificial intelligence that accumulates experience and tracks the situation on the road in order to predict and better organise cargo routes – where, when and how trucks should travel.
Along with the Canadian company Arcadia Laboratories Inc., researchers recently launched a successful project for recognising IT safety incidents using artificial intelligence. The group of VGTU scientists behind this project is the largest team in Lithuania currently working in the area of cyber security. A infiltration detection system is provided with a large quantity of data so that an artificial neural network can learn to detect, based on their actions, new and harmful programming codes and suspicious protocols that might be hiding in the general data stream.
Uniting all of the country’s universities, scientific research institutes, science and technology parks and open-access centres, the open access R&D network Open R&D Lithuania provides more than 2.5 thousand different services in the fields of engineering, IT, biomedicine and biotechnology, material science, physics and chemical technology, natural resources and agriculture. As the largest innovation infrastructure, service and competence network in the Baltic states, run by the Agency for Science, Innovation and Technology (MITA), it facilitates the meeting of Lithuanian researchers developing advanced technologies and entrepreneurs from Lithuania and beyond, encouraging their cooperation. It aids businesses in their search for suitable partnerships with scientists through the Open R&D Lithuania Contact Centre.