Topic title |
Possible scientific supervisors |
Source of funding |
Design and analysis of adaptive classification algorithms
|
prof. dr. Tomas Ruzgas |
state-funded |
Research Topic Summary.
The research goal is to create knowledge about classification algorithms that are effective in the case of heterogeneity. In order to achieve this goal, it is necessary to study mathematical methods based on the statistical analysis of the objects distribution. The goal of this study will be achieved by four main tasks: A) to create a probabilistic model for the distribution of objects and to create model identification procedures; B) to design constructive classification algorithms based on this model and its its identification procedures; C) to offer mathematical methods for taking account of additional information related to object positions and the context between them in classification algorithms; D) to study the proposed solutions, to perform comparative analyzes of alternative classification algorithms on the basis of real data.
|
Adaptive geometric iterative methods for data analysis
|
doc. dr. Svajūnas Sajavičius |
state-funded |
Research Topic Summary.
Many traditional methods for data interpolation and approximation rely on the solution of global linear systems. This feature severely limits the applicability of those methods since even local modifications require a repetitive solution of large-scale linear systems. In recent years, geometric iterative methods have been used increasingly for data interpolation and approximation. These methods have an intuitive geometric meaning, are easily implemented, and allow easy fulfilment of various geometric requirements. Currently, such methods are used in computer-aided geometric design, image and surface reconstruction tasks, reverse engineering, etc. To date, only few studies have been dedicated to investigating the possibilities of applying hierarchical spline technologies in geometric iterative methods. The uniqueness of hierarchical splines is the property of local adaptive refinement. In practical applications, local adaptivity allows reducing the amount of the required computational resources and opens up possibilities for solving much more complex tasks. The main aim of the research will be the development and investigation of geometric iterative methods for data approximation, which will efficiently exploit hierarchical splines technologies. The tasks will include the construction and software implementation of adaptive geometric iterative methods, as well as their theoretical and experimental analysis.
|
Development of algorithms based on physics-informed deep neural network for simulation of dynamical systems
|
doc. dr. Dalia Čalnerytė |
state-funded |
Research Topic Summary.
The ability to collect and process large amounts of data has led to the popularity of machine learning (ML) models for solving various tasks. If enough data is used for training, ML models can identify repetitive structures. This feature allows to use ML models to solve problems that cannot be described by physical equations due to their complexity. Usually, such tasks are solved by applying a simplified system model, but in this case the accuracy of the model deteriorates. The combination of ML and systems described by physical equations would allow modeling the processes of dynamic systems in acceptable resolution, which cannot be used by classical numerical methods due to limitations of computing resources (memory and time).
The goal of the topic - to create an algorithm based on physics-informed neural networks, which would allow modeling dynamic systems with practically acceptable accuracy and limited computing resources.
|
Development of Algorithms for Creating and Real-Time Adjustment of Unmanned Aerial Vehicle Flight Trajectories
|
doc. dr. Andrius Kriščiūnas |
state-funded |
Research Topic Summary.
In recent times, the use of unmanned aerial vehicles (UAVs) is becoming increasingly widespread in both civilian activities, such as field games, transportation, mapping, and military applications, including surveillance and tactical field and internal operations. Due to the extensive use of UAVs, the topic of real-time trajectory adjustments, considering both stationary and dynamic obstacles such as birds or other UAVs, becomes more crucial. Autonomous control of UAV trajectories can be applied to various problem-solving scenarios, such as countering the unauthorized use of UAVs in restricted areas like airports or military facilities. Similarly, automatic trajectory adjustments can be applied in entirely different situations, such as avoiding collisions with other UAVs during cargo transport or remote monitoring.
|
Hierarchicity-based (self-similar) heuristic algorithms for combinatorial optimization problems
|
prof. dr. Alfonsas Misevičius |
state-funded |
Research Topic Summary.
Optimization methods and algorithms are an important constitutive part of the computer science and also artificial intelligence. In particular, the heuristic algorithms play a very significant role, just to mention local search, tabu search, genetic algorithms, (evolutionary) population-based algorithms and their numerous combinations (hybrids). The high efficiency of the combined heuristic algorithms by solving various optimization problems was empirically confirmed several decades ago. The further demand of such algorithms was stimulated by the growth of research and development along with the newly arising tasks. This has become even more actual in recent years.
The current development of heuristic algorithms is continued by not only combining different algorithms, but also by exploiting and expanding the inner structure (architecture) of the algorithms. One of the promising directions in this area is the use of so-called hierarchically-structured (hierarchicity-based) (or simply hierarchical) heuristic (HH) algorithms. The central idea behind the HH algorithms is the multiple utilization (reuse) of the well-known heuristics (like LS, TS, GA). This in connection with what is known as self-similarity - this means that an object (in our case, algorithm) is exactly or approximately similar to constituent parts of itself. The principle of self-similarity is the universal principle; so, we conjecture that for both the computational algorithms, this principle may also be deeply inherent and important. This idea is not very new, and some examples of hierarchical-like algorithms have been already investigated (e.g., iterated/hierarchical local search, master-slave genetic algorithms). Still, this is an area of active and progressing research. Further computational studies are required for accelerating research and to reveal higher potential of the HH algorithms, as well as the new enhancements (hybrid HH algorithms, population-based HH algorithms, etc.).
|
Computer vision algorithms for solving object shape identification problems
|
doc. dr. Armantas Ostreika |
state-funded |
Research Topic Summary.
Computer vision (CV) is an interdisciplinary field that deals with algorithms that enable decision-making by analyzing data from digital images. The goal here is to automate tasks that under normal conditions could only be performed by a human. CV systems are used in many fields, including manufacturing, medical, traffic monitoring, security systems, and more. CV methods allow for automatic analysis of important information when processing large amounts of data. Artificial intelligence algorithms have recently become an integral part of solving computer vision problems.
The proposed topic will aim to improve existing general methodologies focusing on specific production and real-world processes, in order to obtain greater efficiency and reliability, by adapting and improving machine learning methods including Decision Trees, K-Nearest Neighbors, Na?ve Bayes, Support Vector Classifiers, Convolutional Neural Networks. and YOLO algorithm modifications.
Here, CV tasks will include methods of acquisition, processing, analysis and interpretation of digital images; real-world data mining to obtain numerical or symbolic information in the form of decisions. Data interpretation in this context refers to the transformation of digital images into descriptors that can be linked to other action processes and suggest further appropriate solutions.
Focusing on real examples of provided services or production processes, the aim will be to apply and test the obtained algorithms in practice. Examples of applications include CV systems measuring and counting complex shaped products or workpieces, estimating their weight, shape or volume, and inspecting objects at high speed against their predetermined characteristics.
This research is expected to improve existing methodologies and introduce new research ideas based on artificial intelligence and new heuristics methodologies.
|
Metaheuristic approaches for maximally diverse grouping problems.
|
doc. dr. Gintaras Palubeckis |
state-funded |
Research Topic Summary.
Maximally diverse grouping problems (MDGPs) consist of finding a partition of a set of elements into a given number of mutually disjoint groups, while respecting the requirements of group size constraints and diversity. The MDGP models a variety of real-world situations. It is possible to formulate MDGPs with different objective functions and constraints. The purpose is to introduce MDGPs with the Max-Min objective function. Such a variation of the MDGP has several important applications. There is a clear need to develop metaheuristic-based algorithms for solving the Max-Min MDGPs.
|
Detection and parameterisation of ophthalmological structures in eye fundus images
|
doc. dr. Martynas Patašius |
state-funded |
Research Topic Summary.
The eye fundus is one of the few places on the human body where small blood vessels can be observed non-invasively. They are not only evaluated in the diagnosis of ocular diseases, but also in some systemic diseases (hypertension, diabetes). The fundus images also show other structures (e.g. optic disc, drusen, etc.) that are useful for diagnosis. However, even in this field there are still scientific problems that have not been solved. For example, there is no reliable and fast method for tracing the vascular network, which, in turn, would facilitate the determination of various parameters (tortuosity, angles of branching, etc.). This study would seek to address some problems of such type.
|
Design and analysis for statistical hypotheses tests using simulation study
|
prof. dr. Tomas Ruzgas |
state-funded |
Research Topic Summary.
Hypothesis testing is one of the essential branches of data mining, which is based on solving many other tasks (discriminant analysis, image recognition, etc.). The methodology for testing goodness of fit, homogeneity, symmetry and independence hypotheses is gaining more and more attention on newly emerging areas of application: analysis of genetic information processing, analysis of objects of astronomy, analysis of computer technology and its periphery data and etc. Although there are many criteria for checking hypotheses, various authors offer ever new ideas (Nguyen, 2017; Arnastauskaitė, Ruzgas, Bražėnas, 2021). The hypothesis test statistics used in scientific works use a class of probabilistic measurements of many qualities, N-metrics (Klebanov, 2005). The planned research in this dissertation will allow to extend the application of N-metric theory to constructing hypothesis testing statistics. The PhD student of this dissertation have been able to offer N-metric theory based criteria and compare them with some of the classic criteria. Along with the theoretical results, the proposed N-metric type criterion is examine using the Monte Carlo method. To analyze the criteria of homogeneity, symmetry and independence of simple and multiple goodness of fit hypotheses in one-dimensional and multidimensional cases. Examine a wide range of alternative hypotheses. Finally, the algorithms developed should be verified by applying them to real data used in empirical studies (Karpusenkaite, Ruzgas, Denafas, 2016, Milonas, Ruzgas, Venclovas, Jievaltas, Joniau, 2021).
|
Research and implementation blockchain-based reputation model of student
|
prof. dr. Vacius Jusas |
state-funded |
Research Topic Summary.
Blockchain technology enables recording of business transactions online in a secure, transparent, efficient and auditable way. This technology makes it possible to create truly autonomous smart devices that can exchange data (including business data) without the help of a centralized intermediary. Eliminating intermediaries would reduce the costs of many public services, increase the efficiency of the sharing economy and e-government, and could contribute to the implementation of a smart city vision.
Blockchain technology allows the creation of chains of value, i.e. a complete sequence of business processes that enable an enterprise to bring a particular service or product to market and receive remuneration in a safe, efficient, and undeniable manner. An essential element of such chains of value is Smart Contracts, i.e. autonomous agents (programs) operating in blockchains and ensuring that business rules are followed in business transactions. Smart contracts can scan data, perform the necessary calculations according to a specified algorithm and protocol, and pass the results further down to smart contracts in blockchains without human intervention. The practical implementation of blockchain technologies would allow for an innovative transformation of the public services and management sector. One of the small particles of the public sector is university students. After graduation, students obtain a diploma. However, the diploma does not reflect the course of studies and does not provide the employer with enough information about the prospective employee. Therefore, the employer cannot properly evaluate the prospective employee. In the course of the study process, it would be possible to gather information about students' behavior and achievements. Students could supervise the students. The blockchain technology is perfect one for this purpose.
The aim of the research is to apply the blockchain technology to the development of a student’s reputation model.
|
Fractional order differential equations - models and algorithms
|
prof. habil. dr. Minvydas Kazys Ragulskis |
state-funded |
Research Topic Summary.
Fractional order differential equations play an important role in describing natural processes. Memory effects are crucial in the governing equations of such models. Nonlinear chaotic processes taking place in fractional models is a contemporary area of research. Mathematical and computational aspects of modelling will constitute the major part of the planned research program. Expected results include at least several publications in high-level international journals, presentations at international conferences, involvement in activities of research projects.
|