BrainMap

Mr. Liviu-Andrei Scutelnicu

Dr.

- UNIVERSITATEA "ALEXANDRU IOAN CUZA" IASI

Other affiliations

Researcher - ACADEMIA ROMÂNĂ - FILIALA IAŞI (Romania)

Researcher | Teaching staff | Scientific reviewer

Web of Science ResearcherID: ECC-4272-2022

Curriculum Vitae (24/09/2023)

Expertise & keywords

Artificial Intelligence Models (Deep Learning) Applied in the Analysis of Old Romanian Language Call name: P 2 - SP 2.1 - Proiect experimental - demonstrativ
PN-III-P2-2.1-PED-2019-3952 2020 - 2022

Role in this project:

Coordinating institution: ACADEMIA ROMÂNĂ - FILIALA IAŞI

Project partners: ACADEMIA ROMÂNĂ - FILIALA IAŞI (RO); UNIVERSITATEA BUCURESTI (RO)

Affiliation: ACADEMIA ROMÂNĂ - FILIALA IAŞI (RO)

Project website: http://deloro.iit.academiaromana-is.ro/

Abstract:

The project will solve one of the most necessary technology for treasuring in modern conditions the Romanian written heritage. Old documents written in the Cyrillic alphabet will be digitised, cleaned, segmented and then transcribed into the Latin alphabet. In order to do that, state of the art neural and vectorial models will be used and implemented. Proper annotation of primary documents, which will include segmentation of images and contextual transliteration of the content, assured by highly qualified linguist experts, will be used as training data. For a proper organisation of the research activities, we consider documents classified on 3 levels of processing difficulty (low, medium and high) and two types: prints and uncial (hand reproductions of printed majuscules). Manuscripts with ligatures will not be considered in this project. The collection of documents will pass three processing phases: primary (scans of original pages), manual annotation (of columns, lines, words and characters and their Latin transcription - seen as gold files) and the result of automatic processing (test files). Evaluation will be done by comparing test files against gold files. A proposed acceptability threshold should be obtained. The project’s evolution and its acquired data and results will be hosted by a Portal. Significant attention is attributed to dissemination actions.

Resources and technologies for developing human-machine interfaces in Romanian Call name: P 1 - SP 1.2 - Proiecte complexe realizate in consorții CDI
PN-III-P1-1.2-PCCDI-2017-0818 2018 - 2021

Role in this project:

Coordinating institution: INSTITUTUL DE CERCETARI PENTRU INTELIGENTA ARTIFICIALA ,,MIHAI DRAGANESCU''

Project partners: INSTITUTUL DE CERCETARI PENTRU INTELIGENTA ARTIFICIALA ,,MIHAI DRAGANESCU'' (RO); UNIVERSITATEA TEHNICA DIN CLUJ - NAPOCA (RO); UNIVERSITATEA POLITEHNICA DIN BUCURESTI (RO); UNIVERSITATEA "ALEXANDRU IOAN CUZA" IASI (RO)

Affiliation: UNIVERSITATEA "ALEXANDRU IOAN CUZA" IASI (RO)

Project website: http://www.racai.ro/p/reterom/

Abstract:

The ReTeRom complex project brings together, as partners, 4 Romanian institutions with experience in Romanian written/spoken language processing and in the development of language enabled applications. These institutions agreed to achieve knowledge transfer, to reciprocally grant access for own resources and tools and, more importantly, to work together for the unification and improvement of the existing resources and tools, for the development of integrated platforms for acquisition, processing and maintaining of bimodal (text- speech) data. The project will offer to interested third parties a scalable solution for the development of commercial applications.

The main objectives of the ReTeRom complex project are: a) the development of a bimodal aligned corpus (text-speech) of a very high quality and large size – Project 1; b) the development of an integrated and configurable advanced technological platform (machine learning based) which will ensure the standardized processing of the bimodal corpus, processing which will be adapted for the training needs of speech processing systems – Project 2; c) the development of an integrated and configurable interface for automatizing the appropriate annotation of acoustic data. The high quality of automatic annotation of speech raw resources is the only rationale modality to extend in the future the bimodal corpus – Project 3; d) the development of an integrated and configurable platform for high quality and expressivity text-to-speech synthesis (TTS) in Romanian language - Project 4. The high quality of the bimodal corpus will be ensured not only by very accurate tools, but also by a careful validation (and correction, when necessary) by the human experts in the consortium’s teams. The results obtained in the 4 projects will be used for training speech processing systems (ASR and TTS) for the Romanian language. The ASR and TTS will be tested and evaluated on unseen data (not used in the training).

The Contrastive Analysis of Romanian and Spanish Intonation. A Sociolinguistic Approach Call name: Projects for Young Research Teams - RUTE -2014 call
PN-II-RU-TE-2014-4-2298 2015 - 2017

Role in this project:

Coordinating institution: UNIVERSITATEA "ALEXANDRU IOAN CUZA" IASI

Project partners: UNIVERSITATEA "ALEXANDRU IOAN CUZA" IASI (RO)

Affiliation: UNIVERSITATEA "ALEXANDRU IOAN CUZA" IASI (RO)

Project website: http://http.//www.soroes.ro

Abstract:

This research is meant to analyse the state of sociolinguistic research and methodology in Romanian and Spanish intonation. During the last few decades, the interest in describing and modeling intonation has steadily grown, both from a theoretical and an experimental perspective. With this research we aim to fill a gap in the recent literature about the intonation models in both languages, which has largely concentrated on the analysis of read data, by placing particular emphasis on the challenges posed by the analysis of spontaneous speech working with researchers with first-hand experience of most intonation methodology from different foreign universities (University of Oviedo, of Vigo and of Zurich). This research is an approach of sociolinguistic research and methodology in intonation. Although the variationism methodology has long been applied to the description of language variation, it seems that both languages intonation has not developed a very fruitful research. However, it is going through its first tentative stages in Spanish language: Moreno-Fernández 1998; Martín-Butragueño 2006,2011,2014; López-Bobo&Cuevas-Alonso 2014; and some in Romanian: Tiugan 1977, Panaite&Turculeţ 2011.

Although some previous studies, which attempted to explain the intonation from a sociolinguistic perspective, were based on such methodology, in some cases, the techniques used might have been not appropriate.In this context, an attempt at a wider, more comprehensive study seems relevant.

MappingBooks - Let me jump in the book! Call name: Joint Applied Research Projects - PCCA 2013 - call
PN-II-PT-PCCA-2013-4-1878 2014 - 2017

Role in this project:

Coordinating institution: UNIVERSITATEA "ALEXANDRU IOAN CUZA" IASI

Project partners: UNIVERSITATEA "ALEXANDRU IOAN CUZA" IASI (RO); SIVECO ROMANIA SA (RO); UNIVERSITATEA "ŞTEFAN CEL MARE" DIN SUCEAVA (RO)

Affiliation: UNIVERSITATEA "ALEXANDRU IOAN CUZA" IASI (RO)

Project website: http://mappingbooks.uaic.ro/

Abstract:

The project develops a new type of electronic product with a high impact in education and tourism. The main users envisioned are high school pupils/students or tourists visiting Romania. The technology makes heavy use of natural language processing, web cartography, web mapping, mixed reality techniques and ambient intelligence/ubiquitous computing. The project invents a continuous self-sustained activity combining research, education and industry. The language research feeds tools and applications targeting toponyms contained in educational and touristic materials, localising them on 2D and 3D hypermaps and putting them in correlation with the reader’s location and related data. The toponyms can be supplemented with different type of information, diagrams or any graphic related materials. If the school book covers the topic of physical geography, the localisation on a general map will be supplemented with thematic maps (geologic, landform, climate maps), and associated thematic information. As for the tourist guide, up-to-date information regarding the planning of the trip, accommodation, etc., which is not covered by the guide will be displayed on the map.

A number of applications and web services will be built on top of the processed texts, some offered to the data providers with the intention to help them augment the number of consumers of those books. A MappedBook is a book connected with locations/events in the real or virtual world and sensible to the instantaneous location (as seized by the mobile/tablet) of a reader. The information made available could possibly be different depending on the moment and the place the reader happens to be. For instance, if the user reads in the guide about a museum and her/his momentary place happens to be in its proximity, the system will try to find out from the web the museum’s open hours. Then, in case there is still time for a visit, it will signal the user about this possibility. Moreover, with the help of mixed reality technologies, the user’s tablet or smart phone will act as a window through which she/he can see, superposed on an image captured from the real world, the position of different targets.

The providers of primary textual data are media players and editing houses, which, based on written agreements, shall contribute their linguistic resources (electronic forms of manuals and travelling guides) to the partners in this project.

Semantics-driven Syntactic parser (for Romanian) Call name: Projects for Young Research Teams - RUTE -2014 call
PN-II-RU-TE-2014-4-1362 2015 - 2017

Role in this project:

Coordinating institution: INSTITUTUL DE CERCETARI PENTRU INTELIGENTA ARTIFICIALA ,,MIHAI DRAGANESCU''

Project partners: INSTITUTUL DE CERCETARI PENTRU INTELIGENTA ARTIFICIALA ,,MIHAI DRAGANESCU'' (RO)

Affiliation: INSTITUTUL DE CERCETARI PENTRU INTELIGENTA ARTIFICIALA ,,MIHAI DRAGANESCU'' (RO)

Project website: http://dev.racai.ro/ti/wordpress/

Abstract:

The interdisciplinary project “Semantics-driven Syntactic Parser (for Romanian)” aims at using hybrid methods in language processing and at adopting an integrative perspective in the linguistic analysis, in order to create a tool for the automatic syntactic analysis of Romanian. Its added value will be the use of lexico-semantic resources that will improve the results of the deep syntactic analysis. No such hybrid operational parser is described in international research or development papers and, definitely, the Romanian language lacks any kind of wide-coverage parser.

The envisaged results of the project are: (i) an annotation manual for a dependency treebank, (ii) a reference treebank for Romanian, (iii) lexico-semantic resources for syntactic parsing (for Romanian, but transferrable to other languages, when defined at a conceptual level): (a) combinatorial restrictions between words and (semantic) word classes (defined on the Romanian WordNet) accompanied by a reliability score; (b) valency frames for verbs, which contain, for each sense of a verb, the number of arguments and the syntactic and semantic selectional restrictions on their lexicalization; (iv) the hybrid parser.

The project will have a major impact in machine translation, in Romanian corpus annotation, in the creation of strategic resources for Romanian, in the development of applications based on natural language technologies, both by the scientific community and by industry.

FILE DESCRIPTION	DOCUMENT
List of research grants as project coordinator or partner team leader
Significant R&D projects for enterprises, as project manager
R&D activities in enterprises
Peer-review activity for international programs/projects