The project focuses on perspective and ever-expanding field of data processing and extraction of information of valuable content from the data. The methods for information extraction are increasingly using selected methods of artificial intelligence such as neural networks and evolutionary algorithms. Combined models or algorithms for text processing (text mining) come to the fore. The need of these topics is evident from the fact that each year, the volume of data is doubling. At the same time, only 20 % of the available data are processed.
Centralizing media planning to enable agile customer relationships
Program
Operational Program Prague - Pole of Growth Czech Republic
Budeme se podílet na přípravě, návrhu a řešení Systému pro podporu agilního přístupu k zákazníkům prostřednictvím centralizace mediálního plánování, pro firmu MARKETUP
The project is focused on data mining of non-structured data from large-scale sources. The data are represented by textual and image/video information. To porcess the data, selected methods of artificial intelligence such as neural networks/deep learning, evolutionary algorithms and methods of image processing will be used. The need of these topics is evident from the fact that each year, the volume of data is doubling. At the same time, only 20 % of the available data are processed.
Development and testing of algorithms for predictive behavioral analysis of persons crossing the EU's external borders
Program
Program bezpečnostního výzkumu pro potřeby státu 2016 - 2021 (BV III/2 ? VZ)
In order to achieve the objectives of the project, ie the establishment of a system for the early detection of persons crossing the external borders of the EU and the Czech Republic, it is necessary first to prepare available data in a format suitable for the use of data mining methods. Already preparing data and extracting interesting symptoms will be an iterative process because after getting feedback from the customer, the symptoms can be further improved. The data will be processed by several approaches that are appropriate to address this type of problem. The data mining method can be divided into several groups according to the use method, anomaly detection methods in the data, predictive models for revealing repeating patterns, clustering methods for searching similar persons and flights, and other methods that can be beneficial. Due to the nature of the problem, a relatively extensive prototyping and research phase will be necessary before the implementation phase of the production version of the selected algorithms will work. A prerequisite for successful algorithm tuning is the relatively intense synergy of the assignor, whether in terms of feedback to partial results or annotation of data. The research team will use state-of-the-art methods to minimize work on the part of the sponsor - such as semi-supervised or active learning models.
Enhancing capacities for technology transfer and technology uptake in the field of ICT
The project will focus on supporting the development of technology and knowledge transfer in nine countries – the Czech Republic, Germany, Slovakia, Croatia, Serbia, Bosnia and Herzegovina, Montenegro, Romania and Ukraine. The project’s core is the search for new synergies, cooperation and exchange of experience between innovators in each country, while the project will provide support services for developing business plans in the field of technology transfer and commercialization. The project includes creating an incubation/acceleration programme – Danube Digital Accelerator for ICT – enabling 30 participants to learn about perspective technologies. In parallel, the project will enable access and networking of innovative activities and experts in an international environment through the digital tool EXPERTS.AI. Another important part of the project is to map the ICT innovation environment in 4 non-EU partner countries and support the development of their S3 strategies.
The CapTTict project is co-financed by the European Union under the Interreg Danube Region programme.
Fostering Sustainable, Balanced, Equitable, Place-based and Inclusive Development of Rural-Urban Communities' Using Specific Spatial Enhanced Attractiveness Mapping ToolBox
One of the key tasks faced by the rapidly developing area of knowledge discovery in multimedia data, especially from video data, is detecting various human activities. The project will have two objectives, both belonging to basic research: On the one hand, it will contribute to further development of methods for the discovery of different kinds of knowledge in video data recording human activities, such as knowledge concerning shape, motion, color, background, scene, person-object interactions, or thermal infrared images making use of the available experimental environment at the Image Processing Laboratory at FIT CTU. On the other hand, it will perform research into different ways of fusion of multiple classification or regression models in such knowledge discovery tasks. In connection with that objective, the project will rely on the previous experience of its team members from research into combining classifiers and aggregating regression models.
Modern data-mining methods for advanced extraction of information from data
The project focuses on perspective and ever-expanding field of data processing and extraction of information of valuable content from the data. The methods for information extraction are increasingly using selected methods of artificial intelligence such as neural networks and evolutionary algorithms. Combined models or algorithms for text processing (text mining) come to the fore. The need of these topics is evident from the fact that each year, the volume of data is doubling. At the same time, only 20 % of the available data are processed.
New directions in blending evolutionary techniques with data mining
The project focuses on perspective and ever-expanding field of data pre-processing and extraction of information of valuable content from the data. The methods for information extraction are increasingly using selected methods of artificial intelligence such as neural networks and evolutionary algorithms. Combined models or algorithms for text processing (text mining) come to the fore. The need of these topics is evident from the fact that each year, the volume of data is doubling. At the same time, only 20 % of the available data are processed.
Cornerstone of successful data mining is to choose a suitable modelling algorithm for given data. Recent results show, the best performance can be achieved by an efficient combination of diverse models or classifiers. In this project, we develop new algorithms for ensembling models or classifiers. Within our experimental platform FAKE GAME, we can compare new algorithms to already implemented state-of-the-art solutions in the field.
Existuje mnoho algoritmů na optimalizaci dopředných neuronových sítí, většina z nich je však určena pro sítě s jedním typem neuronů. Náš výzkum ukázal, že neuronová síť s různými typy neuronů je universálnější a mnohdy přesnější. V tomto projektu se snažíme vytvořit efektivnější algoritmy pro učení takových druhů neuronových sítí.
Software for evaluation of age with use of pelvis in Retrospective Anthropology, Archeology and Forensic Sciences
Program
Programme of applied research and experimental development in social sciences and humanities ETA
The main objective and output of the project is to create user-friendly software for automatic evaluation of the age of adult skeletal remains. SW will allow age estimation based on sophisticated anthropological and statistical approaches, but its use will be easy even for the general bioarchaeological or forensic public.
The aim of the project is to reduce subjectivity and at the same time to increase the accuracy and reliability of methods for assessing the age of skeletal remains. The approach is based on quantification of changes in the surface area of the pelvic bone. It is an interconnection of several areas of science, from anthropology, archeology and acquisition of 3D data by surface scanners, through their processing by sophisticated statistical approaches, to computer science.
Aplikační garant
Tools for Automatizing the Quality Assurance in Large Business Intelligence Systems and Data Warehouses
Program
Programme of applied research and experimental development ALFA
Environments of BI Systems and DWHs consist of thousands of program artefacts. An automatized management of a large amount of the artefacts requires highly powerful tools, which have been lacking in the market. Using recent theoretical results and long-term practical experiences with a maintenance of environments based on SQL and XML standards, we will create tools that will represent a revolutionary change in the area of the current tools of Quallity Assurance, wherein large financial means have been invested.
