Table of Contents
- Applied Artificial Intelligence Group
- Business Information Technology Group
- Centre for Digital Forensics and Cyber Security
- High-assurance Software Laboratory
- Information Systems Group
- Laboratory for Compositional Systems and Methods
- Laboratory of Language Technology
- Laboratory for Proactive Technologies
- Next Gen Digital State Research Group
- Nonlinear Control Systems Group
Applied Artificial Intelligence Group
The Applied AI Group conducts research in application of AI methods in various fields and systems. We investigate applicability of machine learning, ontology based reasoning, automated theorem provers, knowledge discovery and other AI methods for solving digitalisation problems of different industrial and governmental stakeholders.
Our previous research has been concentrated on building software development methods and tools (e.g. CoCoViLa) with AI components, basically with program synthesis and ontology based knowledge representation components.
During a number of decades several software tools that facilitate AI techniques have been developed by the group. The following is a list of tools that are still in use or under deveopment:
- CoCoViLa – visual model-based software develpment environment;
- WhiteDB – a lightweight NoSQL database library;
- GKC – discussion tool on large knowledgebases.
Currently we work on topics like application of AI methods in spatial data analysis, using machine learning for risk management in e-commerce and for public service delivery. The corresponding projects are listed as follows:
- Applied research for creating a cost-effective interchangeable 3D spatial data infrastructure with survey-grade accuracy
- Applied research for e-commerce EU VAT and duty declaration (as from 2021) digitalisation
- Machine learning and AI powered public service delivery
Business Information Technology Group
The focus of the business information technology group is a theory and practice of dependable, interoperable and evolutionarily changeable enterprise applications and the education of a future specialist in the field.
Centre for Digital Forensics and Cyber Security
TalTech Centre for Digital Forensics and Cyber Security works towards raising Estonian cyber security competence and capacity through education, research and knowledge transfer. The research team includes experts from various scientific disciplines, including computer science, law and psychology. Such a team can take on today’s complex cyber security issues that require an interdisciplinary approach.
The main research directions of the Centre are:
- critical Information Infrastructure Protection (focus on eGovernance and transportation sectors);
- network monitoring;
- digital forensics;
- education research;
- cyber security strategy and policy.
High-assurance Software Laboratory
The group conducts research into theories, methods and tools for developing high-assurance software, specializing on both proofs (certified software) and testing.
- A dataset for detection of Android malware was collected and curated that covers the full history of the operating system.
- A construction was given for a digital signature scheme that makes it possible to use the same key pair multiple times based on timestamping.
- Focused sequent calculi were identified for partially normal skew monoidal, skew prounital closed ja skew symmetric monoidal closed categories.
Information Systems Group
The Information Systems Group conducts research in large- and ultra-large-scale IT systems. We investigate the architecture, design, realization and management of IT system landscapes, high- volume data-intensive systems, high-volume workflow-intensive systems, massively resource-intensive systems and highly distributed systems. In particular, we investigate the next generation of digital government technologies and digital government ecosystems. Together with our partners from industry, academia and the public sector we strive for excellent solutions for non-standard, mission-critical IT system problems
Laboratory for Compositional Systems and Methods
The laboratory started functioning in 2020
The group's goal is to study compositional techniques in the context of models of computation, understood broadly. Compositionality means that syntactic descriptions for (open) systems are designed to be compatible with their semantics. While the examples motivating the research come from a broad section of scientific disciplines (logic, control theory, formal language theory, business processes, game theory, economics, machine learning), we have identified common principles for reasoning about open systems, guided by category theory. These include a semantic universe based on relations rather than functions, and the use of the diagrammatic syntax of string diagrams. String diagrams provide an intuitive calculus for computations via diagrammatic reasoning, and fine-grained control over resources, which is important for faithful descriptions of open systems.
The group's big questions/challenges are:
- design a next generation of programming/specification languages that will be more suited for compositional (and therefore, more trustworthy and reliable) descriptions of systems;
- use compositionality to improve the analysis of systems, including the design of new techniques and algorithms;
- design and implement tools for working with string diagrams, fast-tracking the passage from theory to practice.
Laboratory of Language Technology
The Language Technology Laboratory focuses on the following topics:
- speech recognition;
- speaker, spoken language and accent identification;
- speech corpora;
- phonetics (Estonian language prosody and vocal system, L2 speech);
- various sub-topics of natural language processing.
One of the important activities is the creation of speech technology applications targeted at society as a whole. This includes applications of end-user speech recognition as well as the key integration components that are easy to integrate. Although the focus is on speech recognition in Estonian, most of the software created in the laboratory is not specific to Estonian. The laboratory is a solid open source free software supporter.
Laboratory for Proactive Technologies
The laboratory focuses on theoretical and practical study of networked systems built from stationary and/or mobile software-intensive (proactive) components. Typical components are pervasive computing systems. The research is partitioned into three threads: (1) modelling and verification of situation-aware interaction-centred computation; (2) methods and technologies for acquiring situational information; (3) methods for interpretation of situational information for (proactive) decision making. The long-term goal of the laboratory is the ability to detect and partially control the emergent behaviour in pervasive computing systems.
In addition, ProLab performs research on classification, semantic segmentation and object detection using convolutional neural networks. The methodology has been applied to photographic images, point cloud collections and sound recordings.
Next Gen Digital State Research Group
The Next Gen Digital State (NGDS) research group addresses the technological complexities of how governments can satisfy the current and future needs of their citizens. We focus on digital government ecosystems by investigating technologies that support digital transformation, innovation and implementation.
Our research group collaborates with Estonian and international public sector agencies, ministries, and departments for developing next-generation government-technology through cutting edge research topics focused on artificial intelligence architecture, requirements engineering, data analytics, and understanding the socio-economic effects of technological implementation. We strive to be on the forefront of public sector innovation research!
Nonlinear Control Systems Group
The group is a leading Estonian research unit in automatic control, focusing on nonlinear control systems, including non-smooth, hybrid and time-delay systems. The group has made a significant contributions to the development of constructive algebraic methods and the associated symbolic software package NLControl, which supports research, teaching and applications.
A universal algebraic methodology has been developed that simplifies the study of very different problems for nonlinear control systems from unified perspective. The main idea is to construct sequences of subspaces (or submodules) of differential 1-forms that provide a lot of information about the structural properties of the system. For instance, an event-based resource-aware control method based on the concept of differential flatness has been developed.
Although the group is developing predominantly application-independent general methods determined by the dynamic properties of the mathematical models, we have been recently focused on a few carefully chosen applications, some of them addressed within the joint topics in the Estonian Centre of Excellence on IT, our group is part of. These include control of autonomous underwater vehicles and ionic polymer-metal composite actuators. Within the last few years, special attention has been paid to the study of practical problems arising in limits of renewable energy integration, and determine the possible limitations of distributed energy storage devices in low inertia power systems utilizing methods from optimal control theory.