Projects / Programmes
Advanced methods of interaction in telecommunication
January 1, 2018
- December 31, 2027
| Code |
Science |
Field |
Subfield |
| 2.08.00 |
Engineering sciences and technologies |
Telecommunications |
|
| 1.07.00 |
Natural sciences and mathematics |
Computer intensive methods and applications |
|
| Code |
Science |
Field |
| T180 |
Technological sciences |
Telecommunication engineering |
| Code |
Science |
Field |
| 2.02 |
Engineering and Technology |
Electrical engineering, Electronic engineering, Information engineering |
| 1.01 |
Natural Sciences |
Mathematics |
Multimodal interaction, audio/video content, content-initiated services, automatic speech recognition and synthesis, language resources, speech centered translation, embodied conversational agent, supportive living environments, formal methods, reversible computing, model checking, SDN, SDL, IoT
Data for the last 5 years (citations for the last 10 years) on
April 17, 2024;
A3 for period
2018-2022
| Database |
Linked records |
Citations |
Pure citations |
Average pure citations |
| WoS |
173 |
2,014 |
1,843 |
10.65 |
| Scopus |
334 |
3,801 |
3,342 |
10.01 |
Researchers (21)
Organisations (1)
Abstract
The research work will be conducted within two research areas: multimodal interfaces and ensuring the system correctness. In the area of multimodal interfaces the research will be carried out in the field of audio/video enabling technologies that provide natural and user friendly interaction in modern communication environments, such as intelligent or supportive living environments, in the field of content initiated communication services, as well as large-scale multimedia information management, using audio/video content processing techniques (e.g., within IPTV environment). In the field of acoustic processing the research will focus on development of algorithms that will allow hands free communication. The research in the field of acoustic modelling and automatic speech recognition will address efficient approaches for continuous and conversational speech recognition with emphasis on highly inflectional languages and with focus on different factors, which are present in human speech and have statistically significant influence on speech recognition accuracy. In the field of machine translation the research work will focus on solutions for highly inflected Slavic languages, to which Slovenian language belongs. The work will include analysis of one- and two-pass architectures for machine translation and written and spoken language. The solutions for general and domain specific translation will be analysed and integrated in different application environments. In the field of embodied conversational agents, the proposed research work will tackle the automatic synthesis of personalized and highly natural interaction between man and machine. The work will incorporate interaction as a full life cycle of communication and will address all of its components, from multimodal input processing, understanding of perceived information, to multimodal output planning and conveying the response via multiple modalities. Research in the domain of brain computer interfaces will address the perception and understanding of the visual evoked potentials as a result of the visual stimulation. The research will focus on the design of binocular interfaces, where each eye of the user is visually stimulated separately with both frequency and binocular phase coded symbols, and the responses to a binocular visual excitation on the left and right occipital lobe of the brain are perceived. The work in the framework of system correctness assurance will include the development of new formal methods and the use of new ones for reactive systems with a special emphasis on the ability of reversible computing. We will explore methods for the development and verification of software defined networks and functionalities of newer versions of SDL and support them in the existing algorithms and tools for the creation of the IoT systems' models whose functional correctness will be possible to analyse by the model checking method.
Significance for science
The convergence of the proposed research work provides the ability to create new starting points for developments and guidelines not only towards more intuitive and convenient human machine interaction, but also to increase the level of awareness of electronic systems in regards to their operational environments and management within unified mesh networks in the concept of Internet of things. Expanding the awareness towards a wider context of the user’s environment based on the integration of new sensor technologies and data fusion within the concept of the Internet of things and big data analytics increases the ambient intelligence of smart environments, while setting new guidelines for further development.
The research of spoken language technology algorithms and methods, with proposed, specific, perspective of language characteristics are novel world-wide. It is expected that the outcome significantly improves the speech recognition accuracy and its role in various telecommunication services, such as: intelligent ambience, content processing, natural communication, … Important role will also be given to influence factors, which are reflected in the level of quality of service and quality of experience. Those two are currently one of the most dynamic interdisciplinary research areas in the field of technology. Spoken language use is considerably underinvestigated compared to written use, in Slovene as well as internationally, mainly due to the lack of authentic databases needed for its research. Our studies in the field of language resources will help filling this gap by using state-of-the-art quantitative-qualitative research methods, and by considering the non-linguistic circumstances which are tightly interwoven with spoken language use.
Incorporating more natural, conversation-like behaviour into user interfaces highly improves adaptability and personalization of interfaces. Thus, it improves usability of ever-more technologically complex systems and environments. The key elements of future solutions in the field of multimodal user interfaces will involve social participation, integration of emotional and social features, affect and socialization of computer systems, especially in terms of recognition and understanding of interlocutor’s needs and requirements.
In the area of system correctness assurance, both action-based and state-based approaches have certain advantages and disadvantages. The state-based approach to the specification and reasoning about reversible systems will represent an alternative to the currently dominating action-based approaches. For modern networking technologies, such as software-defined networks and the Internet of things, it is very important that they function correctly and reliably. As these technologies are relatively new, formal methods and tools for them are not well developed yet. The participation of research groups from all around the world, including Slovenia, is, therefore, important for their development. One of the useful tools is our SpinRCP integrated development environment, which is freely available to all users of the Spin verification tool. By introducing new functionalities in SpinRCP, we will contribute to the development of science and greater use of formal methods in the design of modern networks and protocols.
Significance for the country
Speech and language technology is the crucial research area, which is capable to assure the position of Slovenian language in the group of those modern languages, which can follow the challenges of digitalization. Enormous financial investment, mostly in form of spoken language resources, is needed, to be able to provide satisfactory speech recognition results. Thus, Slovenian language hardly stays in touch with those languages, which have larger amount of spoken language resources available. We would like to point out our previous very active work regarding the development of Slovenian language resources, which are available by ELRA/ELDA or provided freely. It is very important to guarantee the position of Slovenian in the group of modern languages with spoken language technology support, because natural way of communication presents one of the key prerequisites needed by the users, to accept modern technologies. As a consequence, the spoken language technology helps to reduce digital divide. One of such application areas is assisted living environment, where natural communication is anticipated and which is going to be even more important in the future, due to demographic change in society.
Speech databases for continues speech recognition contain recordings of authentic spoken use of language. Language represents cultural heritage of a nation. Since it changes in time, such databases represent highly precious archives of cultural heritage for generations to come.
The proposed research work has an extremely positive influence on future behaviour and awareness of wider range of users in regards to sustainable development of modern information society in frames of expansion of supportive living environments. We have to be aware that nowadays encouragement of users towards usage of such services will change in future into commonly usable living environments, which will assure appropriate starting points for the implementation of new products and services.
The involvement in the research in the field of reversibility in the framework of the COST action will help us to have prompt access to the latest knowledge and to personally get in touch with the most prominent experts in this area. The increase in the number of users of our SpinRCP product will widen the network of experts with whom we exchange the knowledge and experiences.
Most important scientific results
Interim report
Most important socioeconomically and culturally relevant results
Interim report
