Navigation
E-Services
Archil Eliashvili Institute of Control Systems of the Georgian Technical University
Email: martsistem@gmail.com, Phone: 231 98 70 / 599 51 21 28 Address: Georgia, Tbilisi, Mindeli str. 10, 0186
About Institute
1347
Publication
420
Conference
312
Project/Grant
23
Intelligence Property
The aim of the institute is to deepen scientific research and improve its quality; To attract young people to the institute and take care of their development; To widespread dissemination and introduction of research results in various fields.
The mission of the Archil Eliashvili Institute of Control Systems of the Georgian Technical University, since its establishment (1956), has been to care for the development of science and to study and process issues relevant to various fields, the research of which is extremely important both scientifically and practically. Currently, scientific research is conducted at the institute in three basic scientific topics: control theory, development of control systems and devices including control processes in energy systems, and informatics.
At present, the scientific research program for 2024-2026 under the Scientific Research Facilitation Program (Program Code 32 05 04)" at the institute has been implemented: “Control theory, identification, optimization and construction of technical systems and devices, modeling of intelligent processes”, which includes five projects: Identification of complex continuous control systems, development and analysis of continuous and discrete optimization algorithms for practical use; Investigation and processing of the remotely controlled technical systems; Automatic translator compilation system development for low-resource languages; Use of artificial intelligence methods to determine the effectiveness of therapeutic drugs; Investigation and processing of the remotely controlled technical systems.
Structural Units
1 Mindia Salukvadze Department of System Identification and Optimal Control 2 Department of Information Transformation issues 3 Department of Language Modelling 4 V. Chavchanidze Department of Artificial Intelligence Problems 5 V. Gomelauri Department of optimization of the power system structure and energy installations
Scientific Equipment
| Title of the equipment/device | Technical characteristics | Date of issue | Exploitation staring year | Usage/application | Purpose of usage/application | Technical condition |
| Oscillograph С1-101, USSR. | Single-channel, frequency range - 0-5 MHz, amplitude - 0.01-300V, power consumption - 8 watts, weight - 1.5 kg, dimensions - (280 х 156 х 69) mm. | 1989 | 1990 | Electronic measurements | Scientific | Active use - in working condition |
| Voltmeter B7 - 34 A | DC voltage measurement range 1 mV-1 kV, measurement limits -0.1 V / 1 V / 10 V / 100 V / 1000 V, DC voltage resistance measurement range - 100 O / 1 KO / 10 KO / 100 KO / 1 M / 10 M, dimensions 480 X 100 X 410, mass 13 kg. | 1988 | 1988 | Electronic measurements | Scientific | Active use - in working condition |
| Oscillograph СК-111, USSR | Conduction band 0-100 mg Hz, lifting time - not more than 3.5 sec, deflection coefficient - 5-10 millivolts / partition, coefficient of expansion - 5-50 milliseconds / partition, screen dimensions - 100 mm X120 mm, power consumption 300 watts, mass 50 kg. | 1985 | 1985 | Electronic measurements | Scientific | Active use - in working condition |
| Turnery screw-cutting machine tool lathe 1E61MT, Ulyanovsk Machine-Building plant / USSR | Production of large diameter and small length details | 1967 | 1967 | Production of large diameter and small length details | Scientific | Active use - in working condition |
| Turnery screw-cutting machine tool lathe 16БО5А, Kirov mashine-tool plant / USSR. | Particularly high-precision lathes are used to cut metric, modular and inch screws. | 1985 | 1985 | Perform class B precision operations | Scientific | Active use - in working condition |
| Turnery screw-cutting machine tool lathe 1K62, / Mashine-tool plant "Red Proletarian" / USSR. | Universal machine tool, used for cutting five types of screws, precision class H, 23 speed mode in the range 12.5-2000 rpm, electric motor power 10 kW, dimensions 3200X1166X1324, weight 3035 kg. | 1960 | 1962 | Perform high precision operations | Scientific | Active use - in working condition |
| Thermographic camera TIS 60+/USA/ FLUKE CORPORATION | Temperature measurement range -20 +400 degrees Celsius, accuracy 2%, F min 46 cm, weight 0.72 kg | 2019 | 2021 | Temperature spectrum measurement | Scientific | Active use - in working condition |
| Thermostat WTHC-1000/ CHINA/ Zhengzhou Wollen Instrument Equipment Co, Ltd | Temperature range from 5 to 100 degrees Celsius, water bath size 15 liters, pump 0-15 l / min, power 1050 watts, weight 14 kg | 2019 | 2021 | Obtain the required temperature of the fluid and maintain it during circulation | Scientific | Active use - in working condition |
| Water distiller YAZD 510 20/ NIGERIA/TECHMEL& TECHMEL INC | Production 5 l / h, power 4.5 kW, weight 7 kg | 2020 | 2021 | Obtaining distilled water for experimental installations | Scientific | Active use - in working condition |
| Voltmeter DM 3068/CHINA/RIGOL TECHNOLOGIES | Dual display, internal memory, USB device, USB host, automatic data processing. Weight 3.2 Kg | 2018 | 2020 | Electronic measurements | Scientific | Active use - in working condition |
| Voltmeter DM 3058/CHINA/RIGOL TECHNOLOGIES | Dual display, internal memory, USB device, USB host, automatic data processing. Weight 2.5 kg | 2018 | 2020 | Electronic measurements | Scientific | Active use - in working condition |
| Water distiller AE-4/ RUSSIA/LIVAM | Production 4 l / h, power 3 kW, weight 6.4 kg | 2018 | 2019 | Obtaining distilled water for experimental installations | Scientific | Active use - in working condition |
| Mixer MM-1000/LATVIA/BIOSAN | Rotational movement 40-1000 turnover/ min, fluid volume 20 l, fluid viscosity 1000-10000 MPa *s, weight 2.4 kg | 2018 | 2019 | For mixing liquids in experimental mixers | Scientific | Active use - in working condition |
| MULTIMETER UT-804 / CHINA / UNI-TREND TECHNOLOGY | Liquid crystal display, data recording device, USB-device, weight 3.4 kg | 2016 | 2016 | Electronic measurements | Scientific | Active use - in working condition |
| MULTIMETER UT-804 / CHINA / UNI-TREND TECHNOLOGY | Liquid crystal display, data recording device, USB-device, weight 3.4 kg | 2016 | 2016 | Electronic measurements | Scientific | Active use - in working condition |
| MULTIMETER UT-804 / CHINA / UNI-TREND TECHNOLOGY | Liquid crystal display, data recording device, USB-device, weight 3.4 kg | 2016 | 2016 | Electronic measurements | Scientific | Active use - in working condition |
| Generator MATRIX MFG 8216 A/Brand MATRIX/China | Manufacturer MATRIX; Frequency measuring range 0.3...3MHz; Amplitude ≥20Vpp @ 50Ω ; Input impedance 1MΩ; Output impedance: 50Ω; Distortion of sinusoidal waveform ≤1% @ 300mHz...200kHz; Weight 2.2kg | 2016 | 2016 | Electronic measurements | Scientific | Active use - in working condition |
| Generator MATRIX MFG 8255 A/Brand MATRIX/China | Manufacturer MATRIX; Frequency measuring range0.5...5MHz; Amplitude ≥20Vpp @ 50Ω ; Input impedance 1MΩ; Output impedance: 50Ω; Distortion of sinusoidal waveform ≤1% @ 300mHz...200kHz; Weight 2.2kg; Dimensions 91x251x291mm | 2016 | 2016 | Electronic measurements | Scientific | Active use - in working condition |
| Projector View Sonic PJD8353S/View Sonic/China | RESOLUTION 1024 x 768; BRIGHTNESS (LUMENS) 3,000 ANSI; LIGHT SOURCE (WATT) 240; DYNAMIC CONTRAST RATIO 15,000:1; FREQUENCY HORIZONTAL 31–100KHz; FREQUENCY VERTICAL 24–120Hz; PHYSICAL WITHOUT STAND (MM) 357 x 231 x 367; | 2015 | 2015 | For Presentation | Scientific | Active use - in working condition |
| Oscillograph SDS 8202/OWEN/China | bandwidth: 200 MHz Max, Number of channels: 2 , color TFT 800×600, diagonal 20 cm , USB device, USB host, dimensions: 340×155×70 mm, weight 1.82 kg | 2016 | 2016 | Electronic measurements | Scientific | Active use - in working condition |
| Oscillograph SDS 8202/OWEN/China | bandwidth: 200 MHz Max, Number of channels: 2 , color TFT 800×600, diagonal 20 cm , USB device, USB host, dimensions: 340×155×70 mm, weight 1.82 kg | 2016 | 2016 | Electronic measurements | Scientific | Active use - in working condition |
| Oscillograph SDS 8202/OWEN/China | bandwidth: 200 MHz Max, Number of channels: 2 , color TFT 800×600, diagonal 20 cm , USB device, USB host, dimensions: 340×155×70 mm, weight 1.82 kg | 2018 | 2018 | Electronic measurements | Scientific | Active use - in working condition |
| Digital Oscillograph OWON SDS7072E/OWEN/hina | Maximum frequency 70 MHz Number of channels 2 ,Sensitivity 2 mV / div ~ 10 V / div Coef, Memory depth 1M screen 20 cm Interfaces USB host | 2017 | 2017 | Electronic measurements | Scientific | Active use - in working condition |
| Digital Oscillograph OWON SDS7072E/OWEN/hina | Maximum frequency 70 MHz Number of channels 2 ,Sensitivity 2 mV / div ~ 10 V / div Coef, Memory depth 1M screen 20 cm Interfaces USB host | 2017 | 2017 | Electronic measurements | Scientific | Active use - in working condition |
Scientific buildings
International Scientific Works
| Project number/ID | Project title | Name of the grant call | Funding organization | Grant budget (total) | Start/end dates | Principal investigator | Key personnel | Project Summary | Detailed description | Achieved results |
|---|---|---|---|---|---|---|---|---|---|---|
| FR-19-3034 | 240000 GEL | 09.03. 2020-09.03. 2023 | "Development of the modern technologies is unimaginable without highly effective energy aggregates. Effectiveness of such equipment as condenser and electric generator cooling systems, metallurgical, chemical, aviation, space equipment etc., heavily depends on extensiveness of heat exchange. Consequently, study and development of the heat transfer intensification methods is of great practical significance. In addition, because heat transfer intensification is closely related with interaction to the structure of the flow, study of the mentioned issues is of great theoretical significance. As it is known, one of the best ways to intensify heat transfer is the method of artificial roughness. Many experimental and theoretical studies are devoted to the intensification of heat transfer under flow conditions in channels (V.Nunner, V.Gomelauri, D.Dippray, R.Sabersky et all). In this direction, many problems have been thoroughly solved, including by the authors of the project. Despite the achieved results, many problems remain unchanged, including the influence of artificial roughness on the heat transfer to falling water film on the surface. Small studies, including the authors of the project, devoted to this problem obviously cannot give a complete picture of the process. At the same time, due to the widespread use of film cooling technique, a large-scale study of the effect of roughness on heat transfer to a flowing film is very significant. In particular, it is necessary to further study the effect on heat transfer of the height of roughness elements, the relationship of the step between the elements of two-dimensional roughness to their height, the combined roughness, etc. The aim of the project is to solve the problems listed above to make a certain contribution to the study of the fundamental problems of non-isothermal turbulent flow around rough surfaces and to strengthen the experimental base for the development and creation of highly efficient power installations. To achieve the goal, the existing experimental setup will be upgraded. Experiments will be conducted to investigate the problems of the heat transfer process of rough surfaces in a wide range of roughness type, geometric parameters, and also Reynolds number, which have not been studied until now. The obtained results will make a significant contribution to solving the fundamental problems of heat transfer of rough surfaces. The developed recommendations based on the obtained results will be used to create highly efficient energy installations. | Ongoing project | Ongoing project | |||||
| FR-21-3509 | The Compiler of the Georgian-English Grammatical Dictionary | Shota Rustaveli National Science Foundation of Georgia | Shota Rustaveli National Science Foundation of Georgia | 180000 GEL | 2022-2025 | Chutkerashvili Anna | Liana Lortkipanidze, Nino Javashvili, Liana Samsonadze, George Chikoidze, Nino Amirezashvili | The intellectualization of search engines in the Internet space significantly increases the speed and quality of the search. Retrieving and storing information from documents is a joint complex process, and word semantics is one of the important factors in this process. WordNet, one of the most common dictionaries of lexical resources used in the field of modern technologies, was created at Princeton University. It is based on the semantics of the word. The successful development of WordNet dictionaries in different countries gave rise to the Georgian project funded by the Shota Rustaveli National Science Foundation of Georgia - "GeWordNet Compiler of the Georgian Word Network". GeWordNet is a large semantic network. It is characterized by two types of semantic relations: lexical (word-word) and conceptual (concept-concept). The most important lexical relation is synonymy. The basic structural unit of the GeWordNet thesaurus is not a single word, but a synonym row, so called, a synset which combines words and concepts of similar meaning. A finite number of associative-semantic relations are established between synsets, such as hyponymy (subtype-general class), meronymy (part-whole), lexical relation (causation, presupposition), etc.; Hyponymy plays the main role among them, which provides the opportunity to organize the hierarchical (tree like) structure of synsets. GeWordNet thesaurus can be used:• To expand the user's query when searching for information through paradigmatically and syntagmatically related words. Such words are, for example, SynSet components, or "verb-actant" type connections, which are needed for context search; • As a dictionary of formal grammars, especially in the determining of the verb valence, exhaustive description of nouns and adjectives; • To compile specialized dictionaries (for example, medical, economic, geographical, biological, etc.); • To compile dictionaries of different dialects and languages; • To resolve the classic task - ambiguity of words by means of syntagmatic relationships of words; • To increase the quality of document filtering and rubricating in software applications for automatic text processing and information search; • To define meaningfully close texts based on hyperonymic relations. In text search and automatic translation systems, the main emphasis was on morphological and syntactic analysis. Semantic analysis is widely used for most languages today. Unlike the morphological and syntactic levels, the area of semantic analysis is not limited to a single word or sentence. It covers the entire text. The deepest and most adequate reflection of the thought using the linguistic means of the natural language can be achieved only by the semantic presentation of the text. The standard methodology for compiling the WordNet thesaurus is based on the analysis of word meanings: the definitional method, the contextual method and the derivation method. The dictionary is represented by four networks, which combine the main parts of speech (nouns, verbs, adjectives and adverbs). Due to the specificity of the Georgian language, we have added the following structural elements to the architecture of GeWordNet: • Semantic-derivative relations for the reflection of the super-paradigm of the Georgian verb;• Description of role relations;• Lexical functions in semantic description. | Ongoing project | Ongoing project |
| AR/84/4-105-14 | Georgian Wordnet Compiler – GeWordNet FR/463/4-105/12 | Applied Research State grants | Shota Rustaveli National Science Foundation of Georgia | 240000 GEL | 2015-2017 | Lorkipanidze Liana | Amirezashvili Nino, Gegechkori Mary, Samsonadze Liana, Chikoidze Giorgi, Chutkerashvili Ana, Javashvili Nino | The intellectualization of search engines in the Internet space significantly increases the speed and quality of the search. Retrieving and storing information from documents is a joint complex process, and word semantics is one of the important factors in this process. WordNet, one of the most common dictionaries of lexical resources used in the field of modern technologies, was created at Princeton University. It is based on the semantics of the word. The successful development of WordNet dictionaries in different countries gave rise to the Georgian project funded by the Shota Rustaveli National Science Foundation of Georgia - "GeWordNet Compiler of the Georgian Word Network". GeWordNet is a large semantic network. It is characterized by two types of semantic relations: lexical (word-word) and conceptual (concept-concept). The most important lexical relation is synonymy. The basic structural unit of the GeWordNet thesaurus is not a single word, but a synonym row, so called, a synset which combines words and concepts of similar meaning. A finite number of associative-semantic relations are established between synsets, such as hyponymy (subtype-general class), meronymy (part-whole), lexical relation (causation, presupposition), etc.; Hyponymy plays the main role among them, which provides the opportunity to organize the hierarchical (tree like) structure of synsets. GeWordNet thesaurus can be used:• To expand the user's query when searching for information through paradigmatically and syntagmatically related words. Such words are, for example, SynSet components, or "verb-actant" type connections, which are needed for context search; • As a dictionary of formal grammars, especially in the determining of the verb valence, exhaustive description of nouns and adjectives; • To compile specialized dictionaries (for example, medical, economic, geographical, biological, etc.); • To compile dictionaries of different dialects and languages; • To resolve the classic task - ambiguity of words by means of syntagmatic relationships of words; • To increase the quality of document filtering and rubricating in software applications for automatic text processing and information search; • To define meaningfully close texts based on hyperonymic relations. In text search and automatic translation systems, the main emphasis was on morphological and syntactic analysis. Semantic analysis is widely used for most languages today. Unlike the morphological and syntactic levels, the area of semantic analysis is not limited to a single word or sentence. It covers the entire text. The deepest and most adequate reflection of the thought using the linguistic means of the natural language can be achieved only by the semantic presentation of the text. The standard methodology for compiling the WordNet thesaurus is based on the analysis of word meanings: the definitional method, the contextual method and the derivation method. The dictionary is represented by four networks, which combine the main parts of speech (nouns, verbs, adjectives and adverbs). Due to the specificity of the Georgian language, we have added the following structural elements to the architecture of GeWordNet: • Semantic-derivative relations for the reflection of the super-paradigm of the Georgian verb;• Description of role relations;• Lexical functions in semantic description. | Manually constructing a WordNet thesaurus is a time-consuming and long process. Automatic methods of obtaining semantic information are interesting. Some of them are established from a large text corpus using statistical methods on the formation of so called semantic vectors. Syntactically and semantically annotated corpus needed for WordNet should be compiled for Georgian language. The corpus will be filled with texts on various topics, such as medicine, sports, various fields of science, etc.; It will be annotated at the morphological, syntactic, semantic level; Vector space models (Vector Space Models) and Generalized Vector Space Model (GVSM (Generalized Vector Space Model)) algorithms for text information processing will be developed. After the software implementation of these algorithms, the method will be tested. The obtained results will be analyzed and the mentioned methods will be modified according to the specifics of the Georgian language. For the implementation of the Georgian version of WordNet, we will also use an approach based on the procedures of automatic compilation of the GeWordNet thesaurus from existing dictionaries of the Georgian language, which will require several stages. In the first stage, using the algorithms and software developed within the project, databases will be filled from both electronic and printed dictionaries. They will be edited and formatted according to the information fields of the lexical unit, which means dividing each lexical unit according to the following values: head word/concept; definition; synonym/synonyms; antonym/antonyms; illustration/illustrations; Matches from another language, etc. The number of fields is determined by each specific dictionary. In the second stage, with the help of the available software we will perform morphological, syntactic and semantic annotation, for the main word/concept of the dictionary; of synonym/synonyms; of antonym/antonyms; and for match/matches fields from another language. In the third stage, using the algorithms and software developed within the project, a prototype is selected for each synonym row of the dictionary and a hyponym tree is built according to the structure of the thesaurus. At the fourth stage, based on the English-Georgian dictionary, GeWordNet thesaurus of the Georgian language and Princeton's WordNet thesaurus, the Georgian dictionary is adapted to English. | The models created within the project reflect only one certain part of the language, for example, a specific morphological event or syntactic representation. Such systems are only designed for limited texts and provide only a partial representation of the language mechanism. Unlike them, in the complete model, all levels of the language are interconnected. Our work is another step towards the computer realization of a complete model of the Georgian language. The proposed approach is innovative, because there is no thesaurus realized within the full volume of the Georgian vocabulary using WordNet technology. The morphological processor of the Georgian language, already implemented by our group, allowed us to enrich the synonym rows automatically with non-lexical forms such as nouns of: property, non-property, origin, purpose, craft, ordinal, distributive, geographical, quality and other nouns. WordNet is a large semantic network. It is characterized by two types of semantic relations: lexical (word-word) and conceptual (concept-concept). The standard methodology for compiling the WordNet thesaurus is based on the analysis of word meanings: the definitional method, the contextual method and the derivation method. The dictionary is represented by four networks, which combine the main parts of speech (nouns, verbs, adjectives and adverbs). Due to the specificity of the Georgian language, we have added the following structural elements to the architecture of GeWordNet: • Semantic-derivative relations for the reflection of the super-paradigm of the Georgian verb; • Description of role relations; • Lexical functions in semantic description. To realize the Georgian version of WordNet, we used an approach based on the procedures of automatic compilation of the GeWordNet thesaurus from the existing dictionaries of the Georgian language. |
| FR/463/4-105/12 | The Full (Morphological, Syntactical, Semantic) Annotation System of Georgian Language Corpora | Fundamental Research State Grants 2021 | Shota Rustaveli National Science Foundation of Georgia | 148680 GEL | 2013 – 2016 | Chikoidze George | Liana Lortkipanidze, Nino Amirezashvili, Elizabeth Dokvadze, Liana Samsonadze, Anna Chutkerashvili, Nino Javashvili, Lamara Margvelani, Mery Gegechkori | The project "Complete (morphological, syntactic, semantic) annotation system of the corpus of the Georgian language" was implemented at the Archil Eliashvili Control Systems Institute of Tbilisi Technical University with the funding of the Rustaveli National Science Foundation (agreement No. 31/65). The project had a leader - L. Lortkipanidze and four main performers: N. Amirezashvili, L. Samsonadze, A. Chutkerashvili, N. Javashvili. The project was ongoing for three years. The goal of the project was to create a software tool, with the help of which we would be able to semi-automatically annotate text corpora at the morphological, syntactic and semantic levels. For this purpose, the work was divided into the following main tasks: 1) obtaining and structuring the texts of the corpus of the Georgian language, 2) meta-annotating the texts of the corpus of the Georgian language, 3) establishing regular superparadigms of the verb, 4) development of a morphological analyzer, 5) development of a syntactic analyzer, 6) development of a semantic analyzer, 7) determination of lexical-grammatical correspondences with modern normative dictionaries and enrichment of morphological and syntactic dictionaries, 8) complete classification of Georgian verb superparadigms and reflection of appropriate semantic information in the dictionary , 9) software implementation of the corpus interactive analyzer, 10) Placing the annotated corpus on the Internet. A morphological, syntactic and semantic analyzer of the Georgian language was developed within the project. As a sub-corpus, on which we tested the complete annotation system of the Georgian language, we selected the prose of the prominent Georgian writer of our time, Otar Chiladze. The collection includes Otar Chiladze's novels: A Man Was Walking along the Road, A Rooster of March, The Iron Theatre, Everyone Who Finds Me, Avelum, The basket. With the help of the annotated corpus within the project, it is possible: • Finding a specific word form and displaying it as a concordance; • Searching for word form according to lemma; • Searching for a group of word forms according to an intermittent or continuous syntagm; • Searching for word forms according to morphological characteristics; • Obtaining various lexical-grammatical statistical data; • Saving selected lines from the concordance in a separate file. The text of the corpus is annotated with morphological, syntactic and semantic markers, which reflect the morphological, syntactic and semantic structure of the writer's language. 655,811 word forms and 97,155 word usages were identified in the corpus. Homonymy is partially removed in the text at all levels. The corpus is placed on the website of the Technical University of Georgia http://geocorpora.gtu.ge/#/texts. For the development of computer linguistics and for maintaining the appropriate level and distribution of the native language, the existence of modern electronic language corpora is a very important and priority means. It offers both the knowledge of the systematic nature of the language (modelling), as well as the reflection, fixation, study of its concrete material created to date, namely, literary monuments, and their use for the research of the language system (building a language model) and practical purposes (translation, dialogue, language teaching computer systems ). In general, to the extent that annotation includes any kind of analytical information about the language of a text, after successful annotation, invaluable material is gathered for building computer models of the language system and for testing various linguistic hypotheses. This, in our opinion, is one of the important results of the project. | The goal of the project was to create a software tool that would enable semi-automatic annotation of text corpora at the morphological, syntactic and semantic levels. For this, the work was divided into eleven main tasks: 1) Acquisition and structuring of Georgian language corpus texts; For the corpus, we selected the novels of the outstanding Georgian writer of our time, Otar Chiladze: A Man Was Walking along the Road, A Rooster of March, The Iron Theatre, Everyone Who Finds Me, Avelum, The basket. Mr. Gia Darsalia, the director of "Arete" publishing house, helped us in obtaining the electronic texts of Otar Chiladze's novels, for which we are very grateful. All texts were converted into Word files and graphometrically processed. Texts were removed from unnecessary gaps (harbs) and empty paragraphs. A single text base was created according to all the works. The texts were written everywhere in the Sylfaen font. The vocabulary of the corpus was connected to the unified text base with the appropriate identification code. Each unique word is associated with an identification code for author, work, chapter/subchapter, paragraph, and sentence identification code. 2) Meta-annotation of Georgian language corpus texts; International meta-annotation standards were studied: TEI (Text Encoding Initiative), CES (Corpus Encoding Standard), CDIF (Corpus Document Interchange Format), XCES (Corpus Encoding Standard for XML) and XML technology. Based on them, the so-called a technology for creating a "text passport" was developed, which involves writing metadata parameters at the beginning of a text document file. All texts have been assigned a corresponding passport. Based on the international standards of meta-annotation, a web application was developed to be placed in the text corpus. A web server was created and the material of the linguistic corpus of one author was placed in its database. 3) Formation of regular superparadigms of the verb; A verb super-paradigm was formed and selected as a set of paradigms derived from the same lexeme, which, in the case of different initial forms, may have different semantic and grammatical structures. Regular and irregular superparadigms were distinguished. Based on them, verbs were classified. 4) Development of morphological analyzer; A desktop application was created, where programs for morphological analysis of literary and authorial languages are implemented as separate software utilities. The words of the corpus were morphologically analyzed. 655,811 word forms and 97,155 word usages were identified in the corpus. There were 40 numerical names represented by numbers, 38 word usages were recorded with errors. From the remaining 97,076 words, homonymy was removed from 84,686 units, 12,390 words are homonymous, 480 words were analyzed approximately, probable markers were assigned to 4,526 words. 5) Development of a syntactic analyzer; The syntactic annotation should reflect the syntactic structure of the individual sentence, according to which the tree of word connections is built. Each word of the sentence represented by the syntactic graph is paired with its dominant (in a monosyllabic sentence, this word itself is dominant). Except for the nominal sentence, in which the first word denoting the object or event is considered to be the main dominant, the predicate is the main node of the syntactic tree - the dominant. Only one member of the sentence is assigned the main dominant marker - D. The rest of the members are dependent members and all are assigned the marker of the type of connection received with their dominant pair. Syntactically analyzed and syntactic features were assigned to 97,155 items. Syntactic homonymy is not removed in the corpus. 6) Development of a semantic analyzer; When assigning semantic tags to texts, one of the requirements is as follows: an important place should be in the area of use of the lemma (source word) corresponding to the word form, that is, an indication of what words the given word matches and what meaningful relationship it has with other words. Initially, the semantic information of the word as one of the zones of the dictionary was placed with the lemmas of the words in so called "the explanatory -combinatorial dictionary". And then, based on it, the texts were semantically tagged. Lexical-semantic information is written for the word in the texts with the following markers of the group: 1. Main characteristic - part of speech; 2. The lexical-semantic information itself; 3. Derivative (word-forming) characteristics. 7) Determination of lexical-grammatical correspondences - reconciliation with modern normative dictionaries and enrichment of morphological and syntactic dictionaries; Morphological annotation standards adopted by international organizations (EAGLES (Expert Advisory Group on Language Engineering Standards), ISLE (International Standards for Language Engineering), LGR (The Leipzig Glossing Rules), ISO (International Standardization Organization)) were studied. Taking them into account, the dictionaries of the morphological processor were filled with new markers. A software utility for semi-automatically assigning standard markers to morphological features has been added to the Morphology Analyzer desktop application. 8) Full classification of Georgian verb superparadigms and reflection of appropriate semantic information in the dictionary; In the lexicon of the morphological analyzer, a reference to the class to which the super-paradigm belongs is written to the head words of the verbs, and there are listed the addresses of units expressing stepwise paradigms, such as causation - CAUS, process (current) - PROC and result (result) - RES. 9) Software realization of the corpus interactive analyzer; The web application eSketch utility was created, with the help of which it is possible to search for each unique word in the context of the sentence of which this word is a member. Also, in the annotated corpus it is possible to manually remove homonymy in the analyzer window, in which the operator selects one of the several answers to distinguish word form. In the web application, the frequency counter of homonymous wordforms and the concordance presentation of annotated wordforms function according to individual works. 10) Placing the annotated corpus on the Internet; The text of all Otar Chiladze's novels is placed in the corpus. Texts are annotated on morphological, syntactic and semantic levels. Manual removal of homonymy continues. 11) Testing The linguistic corpus of the Georgian language based on Otar Chiladze’s novels is posted on the website of the Technical University of Georgia http://geocorpora.gtu.ge/#/texts. Testing is successful. | A morphological, syntactic and semantic analyzer of the Georgian language was developed within the project. As a sub-corpus, on which we tested the complete annotation system of the Georgian language, we selected the prose of the prominent Georgian writer of our time, Otar Chiladze. The collection includes Otar Chiladze's novels: A Man was going down the Road, March Cock, The Iron Theatre, Everyone that findeth Me, Avelum, The basket. With the help of the annotated corpus within the project, it is possible: Finding a specific word form and displaying it as a concordance; Searching for word form according to lemma; Searching for a group of word forms according to an intermittent or continuous syntagm; Searching for word forms according to morphological characteristics; Obtaining various lexical-grammatical statistical data; Saving selected lines from the concordance in a separate file. The text of the corpus is annotated with morphological, syntactic and semantic markers, which reflect the morphological, syntactic and semantic structure of the writer's language. 655,811 word forms and 97,155 word usages were identified in the corpus. Homonymy is partially removed in the text at all levels. The corpus is placed on the website of the Technical University of Georgia http://geocorpora.gtu.ge/#/texts. |
| №104 | Computer support for language teaching (morphological level) | Intra-Institutional Scientific Grants Competition of the Georgian Technical University | Georgian Technical University | 9000 GEL | 06.05 2011 – 31.12.2011 | Lorkipanidze Liana | Amirezashvili Nino, Samsonadze Liana, Javashvili Nino | Language teaching is one of the most important parts of education. On the basis of knowledge of foreign languages, international contacts in cultural, political, economic and generally any social sphere are formed and born. Without this international relationship, it is difficult, practically impossible, to occupy a decent place in the modern civilized world, both for the state as a whole and for its individual representative. To meet these conditions, it is necessary to improve the language teaching process. It is almost impossible to master a language - the most complex and huge natural intellectual system - as a result of taking any full-fledged course. Further use of the acquired knowledge, as a rule, requires additional regular assistance, which is provided by an ordinary dictionary. At the current stage of technology development, many intelligent systems have been created and realized through computers. Computer language systems are important among them, the deep theoretical basis of which is language modeling, that is, such artificial systems that repeat the main aspects of linguistic behavior: language knowledge, use of knowledge for analysis-synthesis of expressions and acquisition of knowledge. The presented system is focused on the practical implementation of knowledge acquisition. Natural language modeling relies on the dictionary, which plays a leading role in the process of language acquisition and application of acquired knowledge. The computerization of the dictionary is used as the starting point for building a computer system for language teaching. Although conventional ("book") dictionaries are very valuable for solving the task, they have two serious drawbacks: lack of information and "passivity". It should be noted that in a regular dictionary, each lexical item is marked with the only "initial" word form of the corresponding paradigm, which, as a rule, is accompanied by insufficient morphological information to construct a complete paradigm, not only for the "ordinary" user, but also for the specialist linguist. To eliminate this shortcoming, a grammatical dictionary is created, the components of which, emerging from any form of the paradigm for each textual word form, ensure finding its corresponding lexical unit. The lexical item is accompanied by appropriate morphological information. Within the framework of the mentioned project, a package of software tools was created, which will help the user to analyze and synthesize the Georgian language both at the level of form production and derivation. The result of the project: 1. The computer program with a trilingual (English-Georgian-Russian) interface will significantly help people interested in learning the Georgian language. 2. Database that can be integrated with Microsoft Office, "Prompt", "ABBYY", "Google", etc. for the full implementation of the Georgian language component in the editors of well-known firms. | It is almost impossible to master a language - the most complex and huge natural intellectual system - as a result of taking any full-fledged course. Further use of the acquired knowledge, as a rule, requires additional regular assistance, which is provided by an ordinary dictionary. At the current stage of technology development, many intelligent systems have been created and realized through computers. Computer language systems are important among them, the deep theoretical basis of which is language modeling, that is, such artificial systems that repeat the main aspects of linguistic behavior: language knowledge, use of knowledge for analysis-synthesis of expressions and acquisition of knowledge. The presented system is focused on the practical implementation of knowledge acquisition. Natural language modeling relies on the dictionary, which plays a leading role in the process of language acquisition and application of acquired knowledge. The computerization of the dictionary is used as the starting point for building a computer system for language teaching. Although conventional ("book") dictionaries are very valuable for solving the task, they have two serious drawbacks: lack of information and "passivity". It should be noted that in a regular dictionary, each lexical item is marked with the only "initial" word form of the corresponding paradigm, which, as a rule, is accompanied by insufficient morphological information to construct a complete paradigm, not only for the "ordinary" user, but also for the specialist linguist. To eliminate this shortcoming, a grammatical dictionary is created, the components of which, emerging from any form of the paradigm for each textual word form, ensure finding its corresponding lexical unit. The lexical item is accompanied by appropriate morphological information. Within the framework of the mentioned project, a package of software tools was created, which will help the user to analyze and synthesize the Georgian language both at the level of form production and derivation. The result of the project: The computer program with a trilingual (English-Georgian-Russian) interface will significantly help people interested in learning the Georgian language. The existing natural language modeling software products required changing the code of the software module itself for each new category word declination or conjugation rule, which was time-consuming due to the colossal number of morphological features of the Georgian language. The presented system uses the software product created by us - GeoTrans, which allowed us to fill in any rules intended for the morphological model of the target language without interfering in the program code. Using it, the morphological system of the Georgian language was reflected in the module of the computer program for teaching the Georgian language and also the computer dictionary base was filled. Both form production and derivation components of language morphology are implemented in the system. We would like to focus on the synthesis/analysis module of Georgian numerical names. The system represents the inflection of numerical nouns with markers that describe the form and morphological features of the numerical noun. It is bidirectional and can be used to analyze a complex numerical name in any case (for ex. One Hundred and Thirty-Five), or to generate the corresponding image represented by markers and numbers (for ex, 135+ Sg+Gen). In the Georgian language, the representation of numbers in numerical names is much more difficult than the corresponding transformations in other languages, e.g. In English, since complex numerical names are declensional in Georgian, numerals are expressed by a twenty-decimal mixed system. The following types of work are performed during the creation of a computer program of computer support for language teaching: 1) The morphological system of the Georgian language is fully reflected in the software module; 2) A computer dictionary database arranged according to the morphological categories of the Georgian language, their classes, subclasses, groups and subgroups has been compiled; 3) A software module was created for quick filling of the dictionary and morphological rules and form checking; 4) A software module was created, which analyzes the word form relying on the morphological system and lexical base of the Georgian language and produces (generates) paradigms according to the lemma (form) of the word; A computer software product for teaching the Georgian language has been created, which can be integrated with Open Office and Microsoft Office software products, which will help the user to learn the Georgian language (at the morphological level). The dictionary database includes 100,000 source words and all the rules for their formation. About 4,000,000 word forms can be identified through our program. If we look at the principle of work of different editors, we can draw a conclusion that they are created mainly based on the statistical method of processing text corpora, which we consider completely unjustified in the case of Georgian due to the scarcity of existing Georgian text corpora. In our opinion, relying only on them will greatly impoverish the computer lexical base of Georgian. We consider preservation of the vocabulary of the Georgian language to be one of the important results of the presented project. Our product is the first interactive program for learning the Georgian language, which can provide the language learner with deep fundamental knowledge about the grammar and vocabulary of the language. With the completion of the mentioned work, another big step was taken towards the ultimate goal of modern computer linguistics: to create a full-fledged automatic translator. The computer translation system is multi-level and very complex. We realized only one of the components of the lowest level (morphology) of this process - a complete morphological processor of the modern Georgian language. Practice has shown that compiling a morphological processor of any language with the help of the GeoTrans system is a simple and fast process. We assume that we will be able to develop a modern English morphological processor and compile a bilingual dictionary (for 100,000 lexical items) within two years. As a result, we will get a Georgian-English and English-Georgian verbatim translator at the morphological level. | Within the framework of the project, a package of software tools was created, which will help the user to analyze and synthesize the Georgian language both at the level of form production and derivation. The result of the project - a computer program with a trilingual (English-Georgian-Russian) interface will significantly help people interested in learning the Georgian language. |
| FR/463/4-105/13 | Automatic explanatory-combinatorial dictionary as the basis of modeling of the Georgian language | Fundamental Research State Grants 2009 | Shota Rustaveli National Science Foundation of Georgia | 150 000 GEL | 2009-2011 | Chikoidze George | Nino Amirezashvili, Elizabeth Dokvadze, Liana Lortkipanidze, Lamara Margvelani, Liana Samsonadze, Nino Javashvili | For the coordinated functioning of the different levels of the language model and for the effective operation of the model as a whole, it is necessary to base it on an explanatory-combinatorial dictionary. The descriptive-combinational dictionary of the Georgian language differs from traditional dictionaries in terms of volume and variety aspect of information. In the automatic combinatorial dictionary, information is arranged according to zones. These are: zones representing word definition, morphological, syntactic and semantic levels. Unlike morphology, the objects of syntax are (regular) constructions of word forms, not individual word forms. General schemes of constructions are the subject of reflection by the syntactic processor (not the dictionary) itself. The concrete verbal realizations of the schemes form an infinite multitude. The only thing that allows us to include a dictionary in this level of functionality is that syntactic constructions always involve some "dominant" member that defines the private face of these constructions. We refer here to only one but the most important case of "dominance", namely the verb. According to the terminology of (Van Wallin et al. 1997), the verb represents the "nucleus" (nucleus) of the upper "layer" (core - "middle", central construction) of the simple sentence structure, the characteristics of which are primarily determined by the character of this "central construction". The structure of the central construction is determined by verb-actant relationships: By the number of actants, their morphological design and semantic interpretation, i.e. semantic roles (Ca - causative, Ag-agent, etc.), which the language assigns to these actants in the context of the given verb (the core of the construction). The structure of a simple sentence includes two important components, the "dominants" of which are the noun (Noun Phrase –NP) and the verb (Core Structure). The first of them (NP) is less dependent on the character of its dominant (substantive); Instead, the second _ "central (chore) construction" structure is determined by the individuality of its dominant verb (Чикоидзе 2007, 2010), namely, by the super-paradigm type to which this verb lexeme belongs. A certain part of the work on the project was devoted to the division of super-paradigms into types and their classification. The already established part of the classification is reflected in the "trial" subset of lexical items. It is very important that the classification is based on the relationship between predicate-role and verb-actant structures, that is, a peculiar syntactic "interface" between semantics and morphology is implemented and, thus, forms the basis of one of the central components of the syntactic level of the model (Amirezashvili et al. 2009). The system of lexical functions (Мельчук 1999) has a versatile potential to ensure the functionality of the model. Some of them (first of all, "substitutions” (Syn, Der, Conv)) are focused on synonymous transformations, which are assigned the greatest role in the scheme of the language model (Мельчук 1999), the correct combinatorics of words (in particular, during synonymous transformations) is provided by the functions Оper, Func, Labor, but their main purpose is the semantic characterization of the corresponding lexeme (i.e. verb). The functions Gener, Mult, Incep, Proper, etc. provide the possibility of semantic signs ("primitives") of the lexeme. The "primitives" needed to describe the semantics of lexemes can be obtained by generalizing and expanding the meaning of some other lexical functions: so, for example, the function Syn can be interpreted in terms of quasi-synonymous rows (Апресян 1997, 2000), which will enrich the set of "primitives" with the different features with which Members of these ranks confront each other. In conclusion, it should be noted that the computer dictionary built as part of the project to create a computer model of the Georgian language provides a lexical basis, which, first of all, is the morphological component of the dictionary, which "through" the lexical units, i.e. lexemes, represents all the paradigms of the Georgian language, that is, the entire multi-million Georgian word form supplies, Thus, it provides the first stage in the case of the direction of the model functioning ("text→content") and the opposite of the functioning ("content→text") – the last stage, i.e. the output. It should be emphasized that the generation of paradigms corresponding to all (100 thousand) lexical units does not simply create the morphological basis of the model, but completely fulfills the task of the morphological component of the model, that is, completely replaces its morphological level. | The direct input/output of the "content ‹—› text" model is the text, that is, the sequence of textual word forms. It is grammar (and first of all - morphology) that systematizes and organizes this multimillion and rather diverse set, without which the computer model would turn into a meaningless list of millions of units. Thus, "reasonable" and "economical" language models are built on the foundation, which is the sum of the dictionary and the morphological processor: the dictionary "feeds" the processor with all the information necessary for its functioning. The dictionary itself is an ordered set of lexical units: each member of the set corresponds to one of the paradigms (not any word form), which narrows the volume of this set from millions to thousands; the "head" of each unit is the "symbol" of the corresponding paradigm: One of its members (word form), stem or root. In the case of a similar direction of the process ("text ‹—› content"), the first step in the analysis of the next word form is the identification of the word form with the "head" of the unit to which it corresponds, that is, the introductory word form must find the unit that reflects the paradigm to which it ultimately belongs to. The lexical units developed within the project are marked with one of the members of the relevant paradigms: in the case of nouns, in the nominative form of the singular number, in the case of verbs, in the infinitive form of the initial verb. This priority is due to the fact that by the generator included in the dictionary itself, the next procedure will be reduced to the identification of the introductory word form with any member of the generated paradigm, while the "generation" way involves the construction of an additional, rather complex morphological processor. Due to the ease of finding an identical form of the input within the paradigm, it can be said that the dictionary constructed when choosing the (generative) option (almost) also fulfills the function of a morphological processor. The morphological generator can be considered as an expression of "morphological knowledge", on the basis of which it is easy to use this knowledge (morphological analysis/synthesis functionality), as well as to fill and expand this knowledge (appropriate dictionary) In the case of the synthetic direction ("content ‹—› text"), at the last stage of the general process of the morphological level, it forms and outputs the word forms, the sequence of which creates the output text. In the case of the generative approach offered by our dictionary, this procedure is reduced to the identification of the member of the generated paradigm whose morphological characteristics match the information received from "above" and, as the final output, to the issuing of its corresponding word form. Thus, by the morphological component, the dictionary with its units represents the entire multi-million set of word forms of the Georgian language, and not just a list of lexemes represented by their "commander" symbols, as is the case in ordinary dictionaries. The dictionary created within the framework of our project can be the basis of the "Grammatical dictionary of the Georgian language", which will provide a great service to both specialists or learners of the Georgian language, as well as ordinary users, interested in the correct use of their native language. Self-explanatory, morphological, syntactic and semantic zones are presented in the descriptive-combinatory dictionary. Each of them has its own tab. From here, the morphological zone is completely processed and 100,000 words of the modern Georgian language are included in it. | The presented system uses the software product created by us - GeoTrans, which allowed us to fill in any rules intended for the morphological model of the target language without interfering in the program code. Using it, the morphological system of the Georgian language was reflected in the module of the computer program for teaching the Georgian language and also the computer dictionary base was filled. Both form production and derivation components of language morphology are implemented in the system. |
| FR/463/4-105/14 | Georgian Computer prompter for disabled persons | Fundamental Research State Grants 2009 | Shota Rustaveli National Science Foundation of Georgia | 150 000 GEL | 2009-2011 | Lorkipanidze Liana | Liana Samsonadze, Lela Nozadze, Nino Amirezashvili, Elisabeth Dokvadze, Lamara Margvelani, Giorgi Chikoidze, Nino Javashvili | The computer prompter implemented within the project is a typing word program called Prophet, which suggests words to the user based on their initial letters, which allows typing fewer letters than usual on the keyboard. Initially, the program was intended for people with severe motor disorders. It allowed them to put in less effort and speed up the writing process. The expectation paid off. The program was also found to be useful for individuals with lexical disorders (dyslexia and a mild form of aphasia). In addition, "Prompter" is useful for any ordinary user, because with its help, the process of typing text on the computer is accelerated. The dictated word is searched in the dictionary and a list of words beginning with the same letter is displayed on the screen, where the order of the words is determined by the word use frequency in the given language. Lexical items can be added to the dictionary, after which it is modified according to the words used by a particular user. In this way, the active dictionary of Prompter is constantly updated, and when printing, unused words are automatically dropped from the dictionary base, and frequently used words are given high priority. Prompter is compatible with all Windows programs. Prompter turned out to be especially convenient for special texts in such fields as: jurisprudence, history, political science, botany, biology, archeology, language theory, etc. This is probably due to the fact that, as is known, every field has its own unique vocabulary and sentence structure. Therefore, when using Prompter, the active lexical bases are refreshed, and the speed of typing texts within the scope of the lexicon increases significantly already when printing the second or third paragraph. Program Prophet Reference system (HELP) messages are realized for Georgian language. The user is provided with information about the program's capabilities and various working modes. The reference system is a set of files used by the Windows system, and the messages in it are implemented in the Georgian language. Calling for help is done from the dialog window of the prophet. Within the framework of the project, a morphological processor of the modern Georgian language was developed and realized by means of the MESLM multilingual expert system of knowledge accumulation, which consists of three main components. These are: verb, noun and other parts of speech. All of them are united by the so-called paradigm cap that includes the general grammatical information of the initial form. General schemes of presentation of paradigms were developed for each part of speech. In all three cases, the acquisition of knowledge by the expert system was carried out according to the same scheme. As planned, the dictionary of source words and rules is filled with 100,000 items. It is possible to demonstrate the synthesis/analysis of any word form (about 24 400 000 word forms) based on the mentioned dictionary. | Prophet is a program that helps those who have problems working with computer texts. The principle is as follows: after typing the first character on the keyboard, the user is given the opportunity to choose the desired word from the grammatically correct forms of the corresponding words provided automatically. The system works in all versions of Windows. There is already a ready-made program that is used for Swedish, English, French, German, Dutch, Norwegian and Danish languages. The Russian version of the Prophet was also completed, in the creation of which members of the Department of Language and Speech Systems of the Archil Eliashvili Control Systems Institute collaborated with Swedish scientists. Based on this experience, a Georgian version of a similar system was created. The system, which we called "Georgian computer prompter", will significantly facilitate the typing of Georgian texts on the computer. For the Georgian version of Prophet, it is necessary to fill in the dictionary bases. For this, it is necessary to develop a morphological processor of the Georgian language, that is, to create such an automatic software module, which, based on the combination of a given lexical unit and morphological characteristics, ensures the construction of the appropriate word form and, on the contrary, the determination of the appropriate initial lexical unit and appropriate morphological characteristics for this or that word form. The software modules of the morphological processor of different languages that existed until now required changing the software code (algorithm) for each new rule intended for the morphological model, which was especially complicated due to the agglutinative nature of the Georgian language and was associated with large time costs. In more modern expert systems, artificial intelligence methods are used, where knowledge about the morphology of the language is accumulated in an interactive mode. It is true that the problem of interference in the program code has been solved, but their use is still quite difficult, because some of them require the knowledge of special operators, while others require the study of high-level languages. The computer prompter implemented within the project is a program that dictates words while typing, which offers the user words based on their initial letters, which allows typing fewer letters than usual on the keyboard. Initially, the program was intended for people with severe motor disorders. It allowed them to put in less effort and speed up the writing process. The expectation paid off. The program was also found to be useful for individuals with lexical disorders (dyslexia and a mild form of aphasia). In addition, "Prompter" is useful for any ordinary user, because with its help, the process of typing text on the computer is accelerated. The dictated word is searched in the dictionary and a list of words beginning with the same letter is displayed on the screen, where the order of the words is determined by the frequency of use of these words in the given language. Lexical items can be added to the dictionary, after which it is modified according to the words used by a particular user. In this way, the active dictionary of Prompter is constantly updated, and when printing, unused words are automatically dropped from the dictionary base, and frequently used words are given high priority. Prompter is compatible with all Windows programs. Prompter turned out to be especially convenient for special texts in such fields as: jurisprudence, history, political science, botany, biology, archeology, language theory, etc. This is probably due to the fact that, as is known, every field has its own unique vocabulary and sentence structure. Therefore, when using Prompter, the active lexical bases are refreshed, and the speed of typing texts within the scope of the lexicon increases significantly already when printing the second or third paragraph. To finish the work of the "Prophet" program, its dialog window should be called. When closing this window, to the question: "The dictionary of subjects has changed, should I save it?” we should answer positively so that our active dictionary would be saved and we need even less effort when typing texts. The program Prophet Reference system (HELP) messages is implemented for the Georgian language. The user is provided with information about the program's capabilities and various working modes. The reference system is a set of files used by the Windows system, the messages are in Georgian. Calling for help is done from the dialog window of the prophet. "Prophet" software is based on five dictionaries obtained by processing Georgian text corpus and based on morphological processor: 1) Main dictionary, which consists of all the words selected from the text corpus (with corresponding frequency); 2) a dictionary of pairs (Pair), which consists of pairs of words selected from the Georgian corpus, which occur one after the other in the text with the corresponding frequency; 3) First dictionary, which consists of 9 words selected according to each letter of the alphabet, which are most often found in the text corpus; 4) Affixes dictionary, which consists of the code and affixes of the corresponding paradigm generating rule of all the words in the Main dictionary; 5) subject dictionary, which is used to store special words combined according to a specific subject or text. Within the framework of the project, a morphological processor of the modern Georgian language was developed and realized by means of the MESLM multilingual expert system of knowledge accumulation, which consists of three main components. These are: verb, noun and other parts of speech. All of them are united by the so-called paradigm cap that includes the general grammatical information of the initial form. General schemes of presentation of paradigms were developed for each part of speech. In all three cases, the acquisition of knowledge by the expert system was carried out according to the same scheme. As planned, the dictionary of source words and rules is filled with 100,000 items. It is possible to demonstrate the synthesis/analysis of any word form (about 24 400 000 word forms) based on the mentioned dictionary. | The main components of Georgian and Russian generative morphology were created, which are needed to build two-directionally morphological processors, that is, such systems, within both morphological analysis and synthesis are carried out: reading content (grammatical features) from the text; Generation of the corresponding word form from the basic (lexical) unit of the given lexical unit. The analysis and synthesis was based on Bf®P (Basic form®Paradigm) type systems developed according to the project, which represent the principle of I.Melchuk's "Content↔Text" model, according to which the dynamic aspect of this model should be represented by a combination of generator and selector. The role of the generator is assigned to the Bf®P systems established here, which, emerging from the basic form (Bf), produce the generation/origin of its entire corresponding paradigm (P). |
| N 6.8.04 | Morphological Level Network Representation and Computer Realization of Generative Grammar | Grants of the of Georgian Academy of Sciences | Georgian Academy of Sciences | 7500 GEL | 2004-2005 | Chikoidze George | Liana Lortkipanidze, Lamara Margvelani, Nino Javashvili | Generative grammar implies the ability to know the language, that the expert of the given language must know the construction of any grammatically correct text and, first of all, the sentence. The sentence construction process is carried out through the stages of direct compilers and transformations. The first of them is represented by tree graphs, and the second by transformation formulas. The morphological component of linguistic knowledge understood from the point of view of native grammar should be the ability to generate paradigms, which, due to the multiplicity and complexity of paradigms, makes the construction and realization of relevant systems a rather serious task, and at the same time confirms the special importance and weight of the morphological component for the global description of Georgian grammar | Generative systems provide an in-depth basis for the development of various automated systems of practical importance, which are implemented through two types of processors. Processors use and realize the knowledge represented by the generative system. Processes of word form synthesis and analysis are meant. These two processes (synthesis and analysis) are the direct basis for almost all automatic language systems of practical importance (translation, dialogue, information search, language teaching, etc.). Combining the originating and generating systems is a very important task both from the theoretical and practical point of view. To successfully solve this rather difficult task, preference was given to the network method, which combines the important aspects of the declarative and procedural aspects of the presentation: visibility and flexibility. Within the framework of the project, the network method of presentation was extended to Georgian and Russian morphology, namely, in the form of a generative system, on the basis of which the construction of generative processors (synthesis/analysis) combined with this system was carried out. | The main components of Georgian and Russian generative morphology were created, which are needed to build two-directionally morphological processors, that is, such systems, within both morphological analysis and synthesis are carried out: reading content (grammatical features) from the text; Generation of the corresponding word form from the basic (lexical) unit of the given lexical unit. The analysis and synthesis was based on Bf®P (Basic form®Paradigm) type systems developed according to the project, which represent the principle of I.Melchuk's "Content↔Text" model, according to which the dynamic aspect of this model should be represented by a combination of generator and selector. The role of the generator is assigned to the Bf®P systems established here, which, emerging from the basic form (Bf), produce the generation/origin of its entire corresponding paradigm (P). |
| GNSF/ST08/3-392 | Identification, control and modeling of complex dynamic systems | Fundamental Research State Grants | Shota Rustaveli National Science Foundation of Georgia | 119000 GEL | 2009-2011 | Besarion Shanshiashvili | Mindia Salukvadze, Givi Bolkvadze, Wilhelm Maisuradze | The project aimed to develop structure and parameter identification, and adaptive control methods and algorithms of complex dynamic systems, which can be described by linear and nonlinear ordinary differential equations accordingly with variable and constant coefficients, as well as stochastic equations. The project also envisaged the investigation of the task of parameter identification of dynamic systems based on the principles of multi-criteria optimization theory. | "In the deterministic case of the project, the tasks of structure and parameter identification of non-linear stationary systems were considered on the set of continuous block-oriented models. The project discussed the problem of parameter identification of linear non-stationary dynamic systems, which was posed as an inverse problem of solving ordinary differential equations with variable coefficients. When investigating the task of multi-criterion parameter identification of dynamic systems, the identification method was developed. In the stochastic case, the task of structure identification of nonlinear dynamic systems was discussed based on the methods proposed by N. Raibman. Parameter identification tasks were solved using recurrent algorithms, namely averaged least squares and Kaczmarzj algorithms. In the project, the problem of adaptive control with parameter identification based on dispersion Hammerstein models was considered for the control system with a closed contour, and an adaptive control system was built. Computer modeling of the developed methods, models, algorithms, and the adaptive control system built on their basis was carried out. Recommendations for applying the developed methods and algorithms to practical objects were developed." | The methods and algorithms of structure and parameter identification and adaptive control of complex dynamic systems described by linear and nonlinear ordinary differential equations, accordingly with variable and constant coefficients, as well as stochastic equations, are processed. The task of parameter identification of dynamic systems is investigated based on the principles of multicriteria optimization theory. |
| Nr 04-77-7067 | Methods and algorithms for automatic determination of the state of complex biomedical objects | An international non-profit association established under Belgian law to promote collaboration with scientists from the former Soviet Union. | INTAS | 9000 Euro | 01.03.2005-01.03.2007 | Manana Khachidze | V. Radzievski, N. Jaliabova | Identification of erythrocytes with pathology based on their image analysis is considered. The classes of erythrocytes that indicate this or that disease are distinguished. Methods and algorithms for solving the problem are offered. | The task of early diagnosis of diseases based on the computer analysis of the image of blood cells - erythrocytes is considered. Image processing of erythrocytes was performed. Signs are distinguished, on the basis of which erythrocytes with normal and different pathologies are identified, that ensures the diagnosis of diseases. Methods and algorithms for solving the problem are offered. | As a result of the conducted research, it was possible to make a computer diagnosis of such diseases as uremia, bone marrow depression, severe forms of hepatitis, some types of anemia and others. |
| GNSF/ST08/7-482 | Investigation of the problems of heat transfer intensification in the turbulent flow and optimization of heat exchangers | Competition for state scientific grants for fundamental research | Shota Rustaveli National Science Foundation of Georgia | 148000 ლარი | 01.01. 2009-01.01. 2012 | Tengiz Magrakvelidze | N.Bantsadze, Kh.Lomidze, J.Rusishvili, N.Lekveishvili | Literary data on the use of mixing devices in technology and existing methods of heat transfer intensification in these devices are analyzed. It is shown that reflective walls are currently used for intensify heat transfer. In addition, it is noted that the use of reflective walls is less effective in terms of increasing the efficiency of devices. In particular, in this case, in conditions of 20% increase in heat transfer, the power required for stirring the liquid increases approximately 5 times or more. It has been shown that the efficiency of using the method of artificial roughness in order to intensify the heat transfer in mixing devices is practically unexplored. In order to investigate the problem, was created an experimental device, on which by the experiments it was determined that in the case of using artificial turbulence, the power required for mixing increases slightly in conditions of a 2-fold increase in heat transfer intensity, which is an undeniable confirmation that the method of artificial roughness is much better than the method of reflective walls. | The development of modern equipment is inextricably linked to the increase in the specific power of energy equipment. And, the creation of such highly efficient devices and their reliable operation are significantly determined by the intensity of the heat transfer processes. In the technological processes taking place in food and chemical industry facilities, mixing volumetric type devices are widely used. The course of technological processes in these devices also greatly depends on the intensity of heat transfer. In such devices, often, along with issues of convective heat transfer intensification, it is relevant to prevent the start of the boiling process under conditions of high thermal loads. Based on all of the above, the processing and study of the methods of heat transfer intensification and reduction of the intensity of sediment formation is of great practical importance. In order to solve the mentioned problems, an experimental unit for studying the heat transfer process during turbulent mixing of a liquid in a large volume was designed and implemented with automated systems of cooling, maintaining stable modes, measurements and control, and providing experimental data processing on a computer. Experiments have been conducted for smooth and rough heat transfer surfaces in a wide range of Reynolds and Prandtl criteria. It has been established experimentally that the creation of artificial roughness on the heat-trannsfer surface leads to a significant intensification of convective heat transfer. It has also been found that the effect of artificial roughness is important at the initial stage of nucleate boiling and is practically negligible in the developed nucleate boiling mode. It has been established experimentally that the power required to stir the liquid in the mixing vessel increases slightly as a result of the creation of two-dimensional roughness on the walls of the vessel. Based on the principles of thermo-hydrodynamic analogy, a calculation equation for the heat transfer coefficient of turbulent flow in channels with high roughness elements is obtained, which qualitatively explains the features of the heat transfer process in such channels. A physical model for heat transfer process in the flow around a rough plate has been developed and a suitable calculation formula for the heat transfer coefficient is obtained, which satisfactorily summarizes the existing experimental data. A physical model of the sediment formation process on porous surfaces has been developed. Based on this model, a calculation formula for sedimentation thickness is obtained, which is in satisfactory agreement with the experimental data. The physical model for the heat transfer process of the roughness surface in the mixing apparatus has been developed. Based on this model the calculation formula for heat transfer is obtained which qualitative agreement with the experimental data. The problem of optimization of the roughness heat-transmitting surface is posed and solved, based on which it is shown that the use of the artificial roughness method is effective in the case when the peaks of the roughness elements are located in the buffer layer of the turbulent flow. | By experiments it was been established that the creation of two-dimensional roughness on the heat-transfer surface in a mixing vessel leads to a significant intensification of convection heat transfer in a wide range of the Prandtl number of the coolant. However, the effect of roughness for a fluid with a large Prandtl number, other things being equal, begins to manifest at smaller values of the Reynolds number than in the case of fluids with a low Prandtl number. The height of the roughness elements (in the studied range) does not affect on the intensity of heat transfer. The maximum intensification of heat transfer by the method of artificial roughness in the mixing device is achieved when the geometric parameter of roughness is s/h = 5. The power required to stir the liquid in the mixing vessel increases slightly as a result of the creation of two-dimensional turbulence on the walls of the vessel. As a result of the experimental research, the curves representing the bubbling boiling process of distilled water on smooth and rough surfaces were obtained in the case of placing the stirrer and the ring tube at the same and different levels. It has been shown that if the stirrer and the heating tube are placed at different levels for a smooth surface, the intensity of heat transfer at the initial stage of boiling is significantly lower than when they are placed at the same level. At the same time, in the case of a hollow heating tube, the intensity of heat transfer practically does not change by changing the levels. The physical model of the heat transfer process of the rough surface in the mixing apparatus has been developed, based on which the calculation formula for heat transfer is in qualitative agreement with the experimental data. Using the thermohydrodynamic analogy, the calculation formula for the heat transfer coefficient during the turbulent flow a rough plate is obtained. A physical model of the sediment formation process on rough surfaces has been developed. Based on this model, a calculation formula for sedimentation thickness is obtained, which is in satisfactory agreement with the experimental data. The problem of optimization of rough heat transfer surface is posed and solved, on the basis of which it is shown that the use of artificial roughness method is effective in the case when the vertices of roughness elements are located in the buffer layer of the turbulent flow and on the outer boundary of this layer. Problems have been identified, the investigation of which is extremely important for determining the mechanism of the heat transfer process for roughness surfaces in turbulent flow. |
| Investigation of the effect of artificial roughness on the intensity of heat transfer of a film flowing down a vertical surface | Intra-institute competition of scientific-theoretical and applied projects | Georgian Technical University | 9250 ლარი | 01.01. 2011-31.12 2011 | Tengiz Magrakvelidze | N.Bantsadze, Kh.Lomidze, A.Mikashavidze | It is noted that in modern heat transfer devices, the heat transfer process can take place on the heat transfer surface under the conditions of a flowing film. Such processes take place in thermal power plant condensers, chemical and technological installations, rocket equipment and others. Since the intensity of heat supply significantly determines the compactness of the mentioned devices, the search for methods of heat supply intensification and the examination of their effectiveness are of great practical importance. The paper expresses the opinion that the method of artificial roughness can be effectively used to intensify the heat transfer. An appropriate experimental device has been created, on which it has been determined by the experiments that the intensity of heat transfer of a surface with file-like roughness is significantly (2-2.5 times) higher than the intensity of heat transfer of a smooth surface. | In modern heat transfer devices, the heat transfer process can be carried out under the conditions of a flowing film on the heat transfer surface. Such processes take place in thermal power plant condensers, chemical and technological installations, rocket equipment and others. Since the intensity of heat supply significantly determines the compactness of the mentioned devices, the search for methods of heat supply intensification and the examination of their effectiveness are of great practical importance. Along with this, since the intensification of heat transfer is related to the impact on the boundary layer of the heat carrier and the change in the structure of the flow caused by this, the study of the mentioned issue is of great theoretical importance. Many experimental and theoretical studies have confirmed that one of the best means of heat transfer intensification is the use of the method of artificial roughness. The effectiveness of this method has been thoroughly studied on the heat transfer of turbulent flow in channels. The influence of roughness on the intensity of heat transfer in the apparatus with a stirrer has been investigated by the executors of the present project in recent years for the first time. In this research cycle, important results were obtained, which were published both in Georgia and abroad. However, many issues still require further study. At the same time, the effect of artificial roughness on the heat transfer intensity of the film flowing on the heat transfer surface is practically unexplored. According to preliminary considerations based on the approximate analysis we have performed, the effect of artificial roughness on the heat transfer of film should be significant, but this needs experimental confirmation. At the same time, if we take into account the widespread use of the flowing film heat transfer process in modern technology, the research provided by the presented project is very relevant. To achieve the set goal, we created an experimental device, which represented an open contour. Experiments were conducted on network water for both smooth and rough surfaces. The experimental area was a tube made of stainless steel, the outer diameter of which was equal to 10 mm. The total length of the test area was 400 mm. Only the lower part of the pipe, the length of which was 200 mm, was heated. The experimental tube was placed vertically. Copper conductors were attached to the ends of the pipe. The experimental area was also equipped with upper and lower camera, in which three unions were located at an angle of 1200 to each other. The inlet and outlet pipes of the heat carrier were connected to the union. The heating pipe of the test area was heated by passing a low-voltage electric current directly into it. The power supplied to the experimental tube was regulated by a regulating transformer. A two-dimensional roughness was created by spirally winding a copper wire on a smooth heat pipe. A thin stainless steel tube with a diameter of 10 mm was used to create a concave surface. The thickness of the tube wall was 0.1 mm. Longitudinal and transverse dimensions of the arrangement of holes on the surface of the pipe were equal to each other and equal to 2 mm. The depth of the pits was 1 mm. According to our estimation, as a result of the creation of grooves on the experimental tube, the surface area of the heat-transmitting tube increased by about 30%. To create a surface with file-like roughness a thick-walled stainless steel tube (D = 10 × 1 mm) has been machined on a lathe by turning a screw groove on the outer surface of the tube. The diameter of the screw was 14 mm, and the diameter of the groove was 0.25 mm. 14 evenly spaced grooves were cut along the entire perimeter of the pipe. After that, by rotating the pipe in the opposite direction, the countersunk screws with the same parameters were cut. After machining the tube in such a way, its surface was formed with closely spaced roughness elements in the shape of a truncated pyramid with a rhombus base. As a result of the creation of file-like roughness, the surface area of the heating pipe has increased by 50%. The combined roughness was created by spirally winding the copper wire on the tubes with grooved and file-like roughness. In the experiments, water consumption, power allocated in the test area (specific heat flow), water temperatures at the entrance and exit to the test area, and the temperature of the test tube wall were measured. Temperatures were measured with chrome-alumel thermocouples. Modern digital tools were used during the measurements. Based on the measured values, it was determined: heat flow, specific heat flow, average temperature of the outer surface of the wall of the heat transfer tube, average temperature of the fluid, average temperature pressure, average coefficient of heat transfer, Nusselt's criterion, Reynolds' criterion. Processing of experimental data was carried out by means of a computer with a program written in Turbo Pascal language. Based on the comparison of the obtained results, the effectiveness of the type of surface roughness was determined. | The analysis of the researches dedicated to the issues of heat transfer intensification by the method of artificial roughness has been made. It has been shown that the effect of artificial roughness on the heat transfer intensity of the film flowing on the heat transfer surface is practically unexplored. An experimental device has been designed and implemented to investigate the heat transfer process of the film flowing on the outer surface of vertically placed smooth and corrugated tubes. It has been experimentally confirmed that the intensity of heat transfer of a surface with file-like roughness is significantly (2-2.5 times) higher than the intensity of heat transfer of a smooth surface. | |
| 7.1.02 | Determining the levels of possibility and optimality of electric power needs satisfaction of Georgia with local energy resources | Grant of the National Academy of Sciences of Georgia | Georgian National Academy of Sciences | 6384 ლარი | 01.01. 2002-31.12 2003 | Tengiz Magrakvelidze | N.Lekveishvili, Kh.Lomidze | In the research the state of the electric energy system of Georgia in 2002 has been analyzed. It has been shown that the system is experiencing an acute crisis, the overcoming of which is essential in the coming years. The energy resources of Georgia have been also analyzed, and based on this, the opinion is expressed that both for sustainable economic development and for ensuring energy security, it is necessary to meet the country's electricity demand mainly using local resources. The mathematical model of Georgian power plants has been processed and the appropriate optimization tasks have been solved. As a result of solving the optimization task, a conclusion was made that Georgia can cover the demand for electricity at the level of 45 billion kWh/year with its own resources. | In the project the state of the electric energy system of Georgia in 2002 has been analyzed. In particular, the following is mentioned: an extremely acute, crisis situation has been created in Georgia. No power plant can develop the design capacity. As a result of all this, the daily output of the existing power plants currently does not exceed 20-22 million kWh. This is when in 1989 the daily production of electricity exceeded 40 million kWh and in addition about 15 million kWh of imported electricity per day was produced. So, even with the level reached in 1989, in the late 1980s, Georgia was experiencing a significant shortage of electricity. One of the main reasons for the above-mentioned crisis situation is the wrong energy policy, which was implemented in Georgia for ten years. For example, in the 80s, the structure of the electric power system was such (according to the installed capacity, it is still the same today) that approximately 60% of the generated electricity came from thermal power plants. This is when only 10% of the technical potential of hydropower resources was utilized. In addition, 90% of the organic heat required for thermal power plants was imported from other countries, and the use of the already scarce local organic heat resources practically decreased to zero. As it is widely known, energy is the basis of the economy, and thus the development of an independent country cannot be imagined with an energy base founded on imports. In order to create a powerful energy base, it is clear that, first of all, it is necessary to restore the design capacities of the existing power plants. At the same time, large-scale works to introduce new capacities into the energy system should be started immediately. In addition, it is necessary to develop such an energy policy, the implementation of which ensures the satisfaction of Georgia's energy needs mainly with local resources. As it is known, mathematical modeling methods have been widely used in recent years to determine the optimal structure of the electric power system. The main goal of the presented work is to determine the possibility and optimality levels of meeting the demand for electricity with the energy resources of Georgia using mathematical modeling methods. During the course of the project, the daily state of the electric energy system of Georgia is analyzed. In particular, the design characteristics of the main power plants (installed capacity, annual output, etc.), current technical condition and capabilities are shown. The energy resources of Georgia are described, both traditional (hydropower resources, organic heating) and non-traditional (wind, sun, geothermal waters and others). In the same chapter, data on international pipelines (Baku-Tbilisi-Ceyhan, Shahdeniz-Tbilisi-Erzurum) are provided. A mathematical model of the electric power system of Georgia is proposed, taking into account the energy resources received in exchange for the laying of both local and the above-mentioned international pipelines. A proper optimization problem is raised and solved. Based on the analysis of the obtained results, appropriate conclusions are made and recommendations are developed. | The research makes recommendations that it is necessary to develop such an energy policy, the implementation of which will give a priority role to the use of local energy resources. In addition, since Georgia has significant hydropower resources, and the supply of organic heating is limited, in the development of the electric power system, a priority role should be given to the utilization of hydropower resources. Hydroelectric power plants should be used both for peak and base (together with thermal power plants) energy. As a result of the solution of the proposed mathematical model and the appropriate optimization problem, it was concluded that Georgia can cover the demand for electricity at the level of 45 billion kWh/year with its own resources. |
Contact
Phone: 231 98 70 / 599 51 21 28
Website: https://msi.gtu.ge/
Email: martsistem@gmail.com
Address: Georgia, Tbilisi, Mindeli str. 10, 0186
Staff
 |
|
|
Manvel KloyaniAcademic Doctor of Science / Researcher |
|
Dali SikharulidzeAcademic Doctor of Science / Researcher |
|
Duduxana CincadzeAcademic Doctor of Science / Researcher |
 |
Besarion ShanshiashviliDoctor of Science / Principal Researcher |
 |
Nino JavashviliAcademic Doctor of Science / Researcher |
|
Ketevan KutkhashviliAcademic Doctor of Science / Researcher |
|
Liana LortkipanidzeAcademic Doctor of Science / Principal Researcher |
|
Nodar GdzelishviliDoctor of Science / Principal Researcher |
 |
Manana JanikashviliDoctor of Science / Researcher |
|
Giorgi GigineishviliDoctor of Science / Senior Researcher |
 |
Nora JaliabovaDoctor of Science / Researcher |
|
|
|
Anna ChutkerashviliAcademic Doctor of Science / Principal Researcher |
 |
Victor KhutsishviliAcademic Doctor of Science / Senior Researcher |
|
Neli KilasoniaAcademic Doctor of Science / Senior Researcher |
|
Ketevan OmiadzeAcademic Doctor of Science / Researcher |
 |
Giorgi ChikoidzeDoctor of Science / Principal Researcher |
 |
Natela AnaniashviliAcademic Doctor of Science / Senior Researcher |
|
Maia MikeladzeAcademic Doctor of Science / Principal Researcher |
|
Nodar MirianashviliDoctor of Science / Principal Researcher |
|
Zaqaria BuachidzeDoctor of Science / Principal Researcher |
|
|
|
|
|
nino amirezashviliDoctor of Science / Researcher |
|
Davit PurtskhvanidzeAcademic Doctor of Science / Principal Researcher |
|
Maia TsertsvadzeAcademic Doctor of Science / Senior Researcher |
|
Tariel KoberidzeDoctor of Science / Researcher |
|
Meri GegechkoriDoctor of Science / Principal Researcher |
|
Aqvsenti MiqashavidzeDoctor of Science / Researcher |
|
|
|
Nodar ShengeliaAcademic Doctor of Science / Researcher |
|
Liana SamsonadzeAcademic Doctor of Science / Researcher |
|
Otar KartvelishviliDoctor of Science / Principal Researcher |
|
Panaiot StavrianidiDoctor of Science / Researcher |
|
|
|
Veriko BakhtadzeAcademic Doctor of Science / Researcher |
 |
Dimitri RadzievskiDoctor of Science / Principal Researcher |
|
Nugzar KavlashviliDoctor of Science / Principal Researcher |
|
Ketevan KvirikashviliAcademic Doctor of Science / Researcher |
 |
Tengiz MagrakvelidzeDoctor of Science / Principal Researcher |
|
Nugzar DadianiAcademic Doctor of Science / Senior Researcher |
|
Irma ArchvadzeAcademic Doctor of Science / Researcher |
|
Levan GvaramadzeAcademic Doctor of Science / Researcher |
|
Mindia SalukvadzeDoctor of Science / Principal Researcher |
|
|
|
Vilhelm MaisuradzeAcademic Doctor of Science / Principal Researcher |
|
Nodar JibladzeDoctor of Science / Principal Researcher |
|
Joseph GogodzeDoctor of Science / Senior Researcher |