Syllabus on the course “Natural Language Processing and Cognitive Modeling”

Approved by Programme Academic Council

Process-verbal 2 from April 10, 2018

Author Vladimir A. Fomichiov

Credits 3

Academic Hours 114

Year of study 1

Mode of study Full-time

1. Pre-requisites: the concept of oriented graph, knowledge of basic ideas of high

school algebra, basic notions of set theory, theory of relational databases, the skills of

developing rather simple algorithms, the skills of systematizing and generalizing in-

formation.

2. Course type: elective

3. Abstract:

The term “natural language” (NL) in linguistics, theory of applied intelligent systems,

data mining and text mining means the collection of the languages used in certain regions by

people for communication, and, in case of many languages, used for writing the books and arti-

cles, in TV and radio programs. The design of natural language processing systems (NLPSs) is

one of the principal branches of the studies aimed at creating applied intelligent systems. The

main attention in this course is paid to setting forth the basic concepts and principal applications

of semantics-oriented NLPSs, i.e., of applied computer systems processing expressions in NL

with respect to the fact that lexical items are associated with one or several meanings and with

respect to a knowledge base representing the knowledge about the considered application do-

main.

Section 1 of the course describes basic concepts of the field and gives an outline of the prin -

cipal application fields. The subject of Section 2 is the principal theoretical approaches with re-

stricted expressive power to formal describing semantic structure of natural language sentences

and discourses: the approaches of first order predicates logic, theory of semantic nets and frames,

conceptual graphs, abstract meaning represenatation.

Section 3 is an introduction to ontologies and Linked Open Data. In particular, the basic ex-

pressive mechanisms of the languages RDF (Resource Description Framework), RDFS, and

OWL (Ontology Web Language) – the basic languages of the Semantic Web project are consid-

ered. The concepts of Deep Web, Linked Open Data are introduced.

Section 4 considers the main approaches to designing NL processing systems for interaction

with ontologies and Linked Open Data.

Section 5 describes the approaches of using NL-interfaces with a restricted natural language

(called a controlled natural language) for forming ontologies.

The subject of Section 6 is the available approaches to designing transformers of knowledge

pieces from an ontology into natural language expressions.

Section 7 describes the principal ideas of designing semantics-oriented NLPSs in business

informatics. In particular, the recommender systems with NL-interfaces and the NLPSs for con-

structing the profile of a person or a company, proceeding from the data dispersed in numerous

textual sources, are considered.

Section 8 sets forth the principal ideas of using natural language processing for semantic

search on the Web.

Section 9 contains an overview of most interesting projects of NLPSs in the quickly pro -

gressing field of bioinformatics. On the one hand, the approaches to designing NL-interfaces to

databases with medical and biological information are analyzed. On the other hand, the projects

aimed at finding in the scientific literature the descriptions of special experiments with proteins

or at discovering the fragments describing the experience of using in practice certain new

medicines or procedures are considered.

Section 10 sets forth the basic ideas of describing in a formal way semantic structures of sen-

tences and discourses in NL by means of the theory of K-representations (knowledge representa-

tions). It is an original theory of designing semantic-syntactic analyzers of NL-texts with broad

use of formal means for representing input, intermediary, and output data. The current configura-

tion of this theory is represented in the monograph by V.A. Fomichov released by Springer in

2010. The theory of K-representations proposes (for the first time in the world) a system of ten

operations on conceptual structures providing the possibility to construct, step by step, semantic

representations of arbitrary complex sentences and discourses in English, Russian, German,

French and other natural languages. The mathematical model describing this system determines,

in particular, a new class of formal languages called SK-languages (standard knowledge lan-

guages). The most general ideas of this model are described. The principles of constructing se-

mantic representations (in other terms, text meaning representations) of questions of many kinds,

statements, and commands, and complex discourses with the help of SK-languages are set forth.

The texts in considered examples pertain, in particular, to business, management, medicine, biol-

ogy, technology, sport.

2

Section 11 sets forth the principal ideas to the approach of the theory of K-representations

to semantic-syntactic analysis of NL-texts. This approach uses original formal means (being un-

derstandable for programmers) for describing the algorithms of semantic-syntactic analysis inde-

pendently on any programming environment. That is why the use of this approach provides good

preconditions for transporting the algorithms from one application domain to other domains and

makes easier the adaptation of the algorithms to new applications. This section describes also the

problem of and the approaches to developing a Semantic Web of a new generation, or Multilin-

gual Semantic Web, or Meanings Understanding Web. Its principal distinguished feature is to be

the well developed computer tools for understanding written texts and oral speech in many natu-

ral languages: English, Russian, etc.

Field of application and normative references

This program of an academic discipline formulates minimal demands to knowledge and

skills of students and determines the content and kinds of lessons and control data.

The program is for lecturers, tutors, teaching assistants, and students of the direction 080500.68

“Business Informatics” of masters’ preparation studying in accordance with the Master Program

“Big Data Systems”.

The program is developed in accordance with:

* The educational standard of the federal state autonomous educational institution of higher pro-

fessional education National Research University Higher School of Economics, level of prepara-

tion: master, confirmed 26.06.2011;

* working academic university plan corresponding to the direction 38.04.05 “Business Informat-

ics” of the masters’ preparation for Master Program “Big Data Systems”, confirmed in 2014.

4. Learning Objectives

The learning objectives are the acquisition of an introductory collection of theoretical

knowledge and methodological foundations in the field of designing and employing natural lan-

guage processing systems (NLPSs), and also practical skills required for program implementa-

tion of NLPSs.

5. Learning Outcomes

As a result of mastering this discipline, a student is:

To know the characteristics of the state of the art of the field of designing and employing

natural language processing systems (NLPSs), the development trends of this field, the principal

classes of NLPSs, the principles of designing main subsystems of NLPSs, the main approaches

to formalizing semantic structure of sentences and discourses in natural language (NL);

3

To be able to represent in a formal way the semantic content of sentences and discourses in

NL by means of (a) first order logic formulas, (b) abstract meaning representation, (b) SK-lan-

guages (standard knowledge languages), determined by the theory of K-representations (knowl-

edge representations);

To possess the skills of reflecting in a formal way the semantics of lexical units (the words

and short word combinations) with the help of SK-languages.

To be able to represent in a formal way the descriptions of various classes of entities (ob-

jects) by means of the Semantic Web project languages RDF and RDFS.

6. Course plan

Theme 1. Basic concepts and applications of natural language processing and cogni-

tive modeling

The notions of natural language (NL), semantics of NL, semantic representation of a NL-

expression, semantic content of a NL-expression. Morphology, syntax, semantics, and pragmat-

ics of NL. Discourses. Referential structure of a NL-text. The role of knowledge about the world

in natural language comprehension. The notions of ontology, natural language processing system

(NLPS), semantic informational technology, linguistic informational technology. The main

classes of NLPSs. Principal objectives of the Semantic Web project. Linked Open Data (LOD).

The query language SPARQL for interaction with LOD. The problem of creating a Multilingual

Semantic Web.

Theme 2. Formal approaches with restricted expressive power to representing se-

mantic structure of natural language texts

Semantic structure of a natural language expression (NL-expression). The notion of se-

mantic role (thematic role, conceptual case, deep case). The notions of an n-ary relation on a

non-empty set, where n ≥ 1. Unary and binary relations. The notion of an n-ary predicate on an

non-empty set, where n ≥ 1. Unary and binary predicates. Logical basises. Two kinds of well-

formed expressions determined by a considered logical basis. The set of terms determined by a

considered logical basis. The interpretation of terms. The set of formulas determined by a con-

sidered logical basis. The interpretation of formulas. The class of first order logic (FOL) lan-

guages. Basic ideas of using the formal means of FOL for constructing semantic representations

of NL-texts. The notion of semantic role (thematic role, conceptual case, deep case). The aim of

semantic role labeling. The use of binary predicates for reflecting semantic roles. Principal re-

strictions of FOL languages from the standpoint of constructing semantic representations of NL-

texts. Basic ideas of the theory of semantic nets and frames. Basic expressional mechanisms of

Abstract Meaning Representation. Semantically-structured informational languages of the Se-

4

mantic Web project. Triplets as the main data structure of Resoaurce Description Framework

(RDF). Examples. Representation of triplets as elementary semantic nets. Basic ideas of Dis-

course Representation Theory. Basic ideas of the Theory of Conceptual Graphs.

Theme 3. Introduction to Ontologies and Linked Open Data

Semantically-structured informational languages of the Semantic Web project: RDF,

RDFS, OWL (Ontology Web Language), OWL2. Triplets as the main data structure of the lan-

guage Resource Description Framework (RDF). Examples. Representation of triplets as elemen-

tary semantic nets. The mechanism of reification in the language RDF. Additional expressive

mechanisms of the language system RDFS. Principal ways of determining classes of objects in

the language for developing ontologies OWL (Ontology Web Language). OWL-based ontolo-

gies. The use of binary relations’ properties in the processing of knowledge stored in OWL-

based ontologies. A short description of additional expressive mechanisms of the language

OWL2. The application of the language systems RDF, RDFS, OWL to the development of on-

tologies and semantic information systems. The Deep Web. The conception of the Linked Open

Data (LOD). State of the art of LOD. DBpedia as a significant part of LOD. The peculiarities of

LOD2.

Theme 4. Natural language processing for interaction with ontologies and Linked

Open Data (LOD)

The problem of convenient interaction with RDF- RDFS- and OWL-based ontologies.

Principal approaches to the design of NL-interfaces to ontologies. The principles of developing

AquaLog - an ontology-portable question-answering system for the Semantic Web. Main con-

structions of the standard RDF query language SPARQL. The peculiarities of the version

SPARQL 1.1. The methods of designing the transformers of NL requests into SPARQL requests.

The query language SQUALL – a controlled natural language as expressive as SPARQL 1.1.

The main ideas of designing a translator from SQUALL to SPARQL. Lexical analysis. Syntactic

and semantic analysis. Generation of SPARQL expressions. Expressive possibilities of SQUALL

compared to SPARQL. The principles of designing a portable NL-interface to ontologies

PANTO. The approaches to cross-lingual natural language querying over the Web of Data. The

main ideas of realizing semantic mapping from NL-questions to OWL queries, according to the

domain-independent framework Three-Phaeses Semantic mapping (TPSM).

Theme 5. Processing of controlled natural language for forming ontologies

The problem of forming and updating an ontology with the help of a natural language in-

terface. The notion of a controlled natural language (CNL). Types of CNLs. History of CNLs.

The ontology editors Gino and Rabbit. Attempto Controlled English (ACE). The main ideas of

5

ACE syntax. The main ideas of ACE semantics. Controlled English existing grammar frame-

works. Principal ideas of the Codeco notation. Codeco applications.

Theme 6. Transformers of knowledge pieces from ontology into natural language

expressions

The problem of getting to know the content of a collection of knowledge pieces on cer-

tain subject stored in an OWL-based ontology. The main ideas of developing the NaturalOWL

system. Content selection. Text planning. Lexicalization. Sentence aggregation. Generation of

referring expressions. Surface realization. Some other approaches to the problem of generating

NL-texts from subsystems of an OWL-based ontology.

Theme 7. Natural language processing systems in business informatics

The principal directions of designing semantics-oriented NLPSs in business informatics,

Types of recommender systems (RecS). User profiling. Content-based recommendation tech-

niques. Collaborative filtering techniques. The advantages of RecS with a natural language inter-

face (NL-interface). The principles of using RecS for intelligent consulting. The main ideas un-

derpinning the architecture and functioning of the RecS Natural Language Assistant (NLA).

Recommender systems for e-tourism. The aims of development and examples of business rules.

Recommender systems for TV users. The use of formal means for the design of RecS with a NL-

interface. RecS for intelligent interaction with bank clients. RecS of other kinds. The approaches

to constructing a profile of a person or a company by an applied intelligent system, proceeding

from the data dispersed in numerous textual sources.

Theme 8. Natural language processing for semantic search on the Web

Challenges of NL processing of Web search queries. Question answering systems. Ques-

tion answering phases. Deep question answering. Shallow semantic structures for constructing

text representations. Answer reranking. The approaches to constructing semantic expansions of

search queries. Semantic matching. Semantic annotations of informational sources and their use

in Web search. The concept of a mashup. Mashups for Web search engines. Wikipedia. The DB-

pedia ontology. The mashups that are already using DBpedia. The principal features of Sz-

takipedia project.

Theme 9. Applications of natural language processing in biomedical informatics

Section 9 contains an overview of most interesting projects of NL processing systems in

the quickly progressing field of biomedical informatics. On the one hand, the approaches to de-

signing NL-interfaces to databases with medical and biological information are analyzed. On the

other hand, the projects aimed at finding in the scientific literature the descriptions of special ex-

6

periments with proteins or at discovering the fragments describing the experience of using in

practice certain new medicines or procedures are considered.

Theme 10. The principles of describing semantic structures in the theory of K-repre-

sentations

This section sets forth the basic ideas of describing in a formal way semantic structures of

sentences and discourses in NL by means of the theory of K-representations (knowledge repre-

sentations). It is an original theory of designing semantic-syntactic analyzers of NL-texts with

broad use of formal means for representing input, intermediary, and output data. The current

configuration of this theory is represented in the monograph [1].

The theory of K-representations proposes (for the first time in the world) a system of ten

operations on conceptual structures providing the possibility to construct, step by step, semantic

representations of arbitrary complex sentences and discourses in English, Russian, German,

French and other natural languages. The mathematical model describing this system determines,

in particular, a new class of formal languages called SK-languages (standard knowledge lan-

guages). The most general ideas of this model are described. The principles of constructing se-

mantic representations (in other terms, text meaning representations) of questions of many kinds,

statements, and commands, and complex discourses with the help of SK-languages are set forth.

The texts in considered examples pertain, in particular, to business, management, medicine, biol-

ogy, technology, sport. The advantages of the theory of K-representations in comparison with se-

mantic nets theory, theory of conceptual graphs, first order predicates logic, and episodic logic

are analyzed.

Theme 11. The approach of the theory of K-representations to semantic-syntactic

processing of sentences and discourses

This section sets forth the principal ideas of the K-representations theory’s approach to

semantic-syntactic analysis of NL-texts. This approach uses original formal means (being under-

standable for programmers) for describing the algorithms of semantic-syntactic analysis indepen-

dently on any programming environment. That is why the use of this approach provides good

preconditions for transporting the algorithms from one application domain to other domains and

makes easier the adaptation of the algorithms to new applications. The logical structure of princi-

pal semantic-syntactic components of a linguistic database is described by means of the follow-

ing notions: a lexico-semantic dictionary, a dictionary of semantic-syntactic verbal-prepositional

frames, and a dictionary of semantic-syntactic prepositional frames. A number of intermediary

data structures are introduced, in particular, a morphological representation (MSSR) of a NL-

text, a classifying representation of a NL-text, and a matrix semantic-syntactic representation of

7

a NL-text. The principles of classifying the input NL texts, semantic role labeling, constructing a

MSSR of an input NL-text, and of semantic assembly of a K-representation of an input NL-text,

proceeding from its MSSR.

Finally, this section describes the problem of and the approaches to developing a Semantic Web

of a new generation, or Multilingual Semantic Web, or Meanings Understanding Web. Its princi-

pal distinguished feature is to be the well developed computer tools for understanding written

texts and oral speech in many natural languages: English, Russian, etc. An original strategy of

transforming the existing Web into a Semantic Web of a new generation is set forth; this strat-

egy was proposed by V.A. Fomichov. The central idea of this strategy is the use of one formal

environment (the SK-languages) for represented the conceptual structures (or the meanings) as-

sociated with various Web-based sources: natural language texts, visual images, films, the mes-

sages sent by computer intelligent agents in multi-agent systems, semantic annotations of Web-

based sources, etc

7. Reading List

Required

1. Banarescu, L., Bonial, C., Cai, S., Georgescu, M., Griffitt, K., Hermjakob, U., Knight,

K., Koehn, P., Palmer, M., Schneider, N. (2015). Abstract Meaning Representation

(AMR) 1.2.2 Specification; github.com/amrisi/amr-guidelines/blob/master/amr.md

2.  Fomichov, V.A.:SK-languages as a Powerful and Flexible Semantic Formalism for the

Systems of Cross-Lingual Intelligent Information Access // Informatica. An International

Journal of Computing and Informatics. 2017. Vol. 41. No. 2. P. 221-232.

3. Fomichov, V.A. Reader on Representing Information by Means of First Order Logic.

Moscow, National Research University Higher School of Economics, 2018 (this source is

distributed via e-mail).

4. Auer, S., Bryl, V., Tramp, S. (Eds.). Linked Open Data – Creating Knowledge Out of In-

terlinked Data. Results of the LOD2 Project. Springer, Lecture Notes in Computer Sci-

ence, Vol. 8661, 2014 – 215 p.

5. Ceri, S., Bozzon, A., Brambilla, M., Della Valle, E., Fraternali, P., and Quarteroni, S.

Web Information Retrieval. Springer: Heidelberg, New York, Dordrecht, London, 2013.

– 284 p.

8

6. Metais, E., Roche, M., Teisseire, M. (Eds.). Natural Language Processing and

Information Systems. 19th Intern. Conference on Applications of Natural Language to

Information Systems, NLDB 2014, Montpellier, France, June 18-20, 2014, Proceedings.

Springer International Publishing Switzerland, 2014, Vol. LNCS 8455 - 268 p.

7. Decker, H., Lhotská, L., Link, S., Spies, M., Wagner, R.R. (Eds.): Database and Expert

Systems Applications, 25th International Conference, DEXA 2014, Munich, Germany,

September 1-4, 2014, Part I, Proceedings, LNCS 8644, Part II, Proceedings, LNCS 8645,

Springer International Publishing Switzerland, 2014.

Optional

8. Sharp, B.,  Sedes, F.,  Lubaszewski, W. Cognitive Approach to Natural Language Pro-

cessing. Elsevier, 2017. - 234 p.

9. Prince, V. and Roche, M. Information Retrieval in Biomedicine: Natural Language Pro-

cessing for Knowledge Integration. IGI Global, 2009.

10. OWL Web Ontology Language Overview , Deborah L. McGuinness and Frank van

Harmelen, Editors. W3C Recommendation, 10 February 2004 (2004)\\ \url{http://

www.w3.org/TR/2004/REC-owl-features-20040210/. Latest version available at http://

www.w3.org/TR/owl-features/}

11. Endres-Niggemeyer, B. (Editor). Semantic Mashups. Intelligent Reuse of Web Re-

sources. Springer: Heidelberg, New York, Dordrecht, London, 2013. – 382 p.

12. Metais, E., Meziane, F., Saraee, M., Sugumaran, V., Vadera, S. (Eds.), Natural Language

Processing and Information Systems. 18th Intern. Conference on Applications of Natural

Language to Information Systems, NLDB 2013, Salford, UK, June 2013, Proceedings.

Springer, Lecture Notes in Computer Science, 2013, Vol. LNCS 7934 - 426 p.

13. Fomichov, V.A. Semantics-Oriented Natural Language Processing: Mathematical Mod-

els and Algorithms. IFSR International Series on Systems Science and Engineering, Vol.

27; Springer: New York, Dordrecht, Heidelberg, London, 2010. -354 p.;http://

www.springer.com/math/applications/book/978-0-387-72924-4

14. Fomichov, V.A. (Guest Editor). Special Issue on Semantic Informational Technologies.

Informatica. An International Journal of Computing and Informatics (Slovenia), 2010.

Vol. 34. No. 3.

15. Fomichov, V.A.: Theory of K-representations as a Comprehensive Formal Framework

for Developing a Multilingual Semantic Web. Informatica. An International Journal of

Computing and Informatics (Slovenia), 2010, Vol. 34, No. 3, p. 387-396; http://www.in-

formatica.si. 

9

16. Fomichov, V.A., Kirillov, A.V.: A Formal Model for Constructing Semantic Expansions

of the Search Requests About the Achievemnets and Failures. In: Allan Ramsay,

Gennady Agre (Eds.), Artificail Intelligence: Methodology, Systems, and Applications.

15the Intern. Conference, AIMSA 2012, Varna, Bulgaria, September 2012, Proceedings,

LNAI 7557, Springer, pp. 296-304.

17. Hladky, D. And Maltseva, S.V. Linked Data Paradigm for Enterprises: Information Inte-

gration and Value Chain. Business Informatics. Scientific-practical journal of National

Research University Higher School of Economics, 2013, No. 2 (24), p. 3-12.

18. Ramsay, A. and Agre, G. (Eds.), Artificial Intelligence: Methodology, Systems, and Ap-

plications, 15th International Conference, AIMSA 2012, Varna, Bulgaria, September

2012, Proceedings. Springer, Lecture Notes in Artificial Intelligence, 2012, Vol. LNAI

7557.

19. Fomichov, V.A. The prospects revealed by the theory of K-representations for bioinfor-

matics and Semantic Web // TALN 2011. RECITAL 2011. Actes de la 18e conference

sur le Traitement Automatique des Langues Naturels. Actes de la 15e Rencontre des Etu-

diants Cercheurs en Informatique pour le Traitement Automatique des Langues. Vol. 1

(Actes: articles long). Montpellier: AVL Diffusion, 2011. P. 5-20.

20. Serdyukov, P., Braslavski, P., Kuznetsov, S.O., Kamps, J., Rüger, S.M., Agichtein, E.,

Segalovich, I., Yilmaz, E. (Eds.), Proc. 35th European Conference on Information Re-

trieval (ECIR 2013): Advances in Information Retrieval. Springer, Lecture Notes in

Computer Science, Vol. LNCS 7814, 2013.

21. Androutsopoulos, I., Lampouras, G., Galanis, D. Generating natural language descrip-

tions from OWL ontologies: the NaturalOWL System // Journal of Artificial Intelligence

Research, 2013, Vol. 48, p. 671-715.

22. Stevens, R., Malone, J., Williams, S., Power, R., Third, A. Automating generation of tex-

tual class definitions from OWL to English // Journal of Biomedical Semantics, 2011,

Vol. 2 (Suppl 2), S5; http://www.jbiomedsem.com/content/2/S2/S5

23. Fliedl, G., Kop, C., Voehringer, J. From OWL class and property labels to human under-

standable natural language // Z. Kedad et al. (Eds.) Natural Language Processing and In-

formation Systems. 12th International Conference on Applications of Natural Language to

Information Systems, NLDB 2007, Proceedings. Springer, Vol. LNCS 4592, p. 156-167

8 Grading System

Kind of control

Form of con-trol

1 year Parameters 3 4

10

CurrentHome Work

Synopsis+

+

Assessment of Home Work -1 week

The volume 12 – 15 printed pages (Times New Roman, 14 pt), assess-ment of results - 2 weeks

Final Exam+

Written test, 90 minutes, assessment of results -1 week

9. Guidelines for Knowledge Assessment

The course is studied during third and fourth modules. A current control is scheduled for

each of these modules. Fourth module includes the final control – an exam. The 10-balls grading

scale is used for the marks in all forms of current control.

The formation of a final mark for the discipline is done as follows.

Formation of accumulated grade for 3 module

During auditorium lessons, the activity of students at lectures and practical lessons, the

participation in discussions, the correctness of solutions to formulated tasks is assessed.

An accumulated grade for current control takes into account the correctness of written Home

Work (indicated in a Working Learning Plan) fulfilled by a student:

Ocurrent = Qhome-work .

The accumulated grade (10-balls scale) for the work at practical lessons - Оauditorium.

A self-study work of students is also assessed: the correctness of prepared home works (the tasks

are given at practical lessons; the completeness of explicating the themes of reports.

The accumulated grade (10-balls scale) for self-study work – Оself-study-work.

The accumulated grade (10-balls scale) for 3rd module is calculated as follows:

Оaccumulated for 3 module = 0,6· Ocurrent + 0,2· Оself-study-work + 0,2· Оauditorium .

The arithmetical manner of rounding the accumulated grade is used.

Formation of accumulated grade for 4 module

During auditorium lessons, the activity of students at lectures and practical lessons, the

participation in discussions, the correctness of solutions to formulated tasks is assessed.

11

An accumulated grade for current control takes into account the correctness of written synopsis

(indicated in a Working Learning Plan) prepared by a student:

Ocurrent = Qsynopsis .

The accumulated grade (10-balls scale) for the work at practical lessons - Оauditorium.

A self-study work of students is also assessed: the correctness of prepared home works (the tasks

are given at practical lessons; the completeness of explicating the themes of reports.

The accumulated grade (10-balls scale) for self-study work – Оself-study-work.

The accumulated grade (10-balls scale) for 4th module is calculated as follows:

Оaccumulated for 4 module = 0,6· Ocurrent + 0,2· Оself-study-work +0,2· Оauditorium .

The arithmetical manner of rounding the accumulated grade is used.

Formation of final grade for the discipline

The final accumulated grade for the discipline is formed as follows:

Оfinal-accumulated = 0,5 *(Оaccumulated for 3 module + Оaccumulated for 4 module).

The arithmetical manner of rounding the final accumulated grade is used.

The final grade for the discipline is calculated in accordance with the following formula:

Оfinal = 0,6·Оexam + 0,4· Оfinal-accumulated

here Оexam – the grade for final exam (a result obtained directly during a final exam).

The grade for final control (a result obtained directly during a final exam) Оexam is of blocking

character, in case of an unsatisfactory grade it is equal to the final resulting grade for the disci-

pline.

The arithmetical manner of rounding the final resulting grade is used.

It is possible that during the final exam a student receives an additional question (an addi-

tional theoretical task), a maximal grade for an answer (or a solution) is 1 ball.

The themes of practical lessons (seminars) and self-studies

Studying basic functions of the text analytics system Cogito Intelligence API.

Constructing a semantic representation of a short sentence or discourse in natural lan-

guage as a formula of first order logic.

Constructing a semantic representation of a sentence in natural language as a graph of ab-

stract meaning representation.

Studying the linguistic database Propbank.

Studying the semantic parsers using Abstract Meaning Representation.

Studying the expressive mechanisms of the Semantic Web languages RDF and RDFS.

Studying the inference mechanisms of the Semantic Web language system RDFS.

Mastering the computer systems for building simple RDF and RDFS based ontologies.

12

Studying the expressive mechanisms of the search requests language SPARQL, devel-

oped for interaction of end users with Linked Open System (LOD).

Studying the linguistic database FrameNet, containing semantic frames of lexical units.

Studying the shallow semantic Shalmaneser, based on the linguistic database FrameNet.

Studying the shallow semantic LTH, based on the linguistic database FrameNet.

Studying basic functions of the semantic parser SEMAFOR, based on the linguistic data-

base FrameNet.

Studying basic functions of Natural Language Commander Two - a natural language in-

terface of a file manager.

Studying the approaches to the formation of ontology with the help of natural language

processing systems.

Studying the approaches to designing transformers of the collections of OWL-expres-

sions (representing the pieces of knowledge) into a natural language text.

The themes of home work tasks

Theme 1: The use of first order logic formulas for constructing a semantic representation of a

short sentence or discourse in natural language (NL)

Theme 2: The use of Abstract Meaning Representation for reflecting the semantic content of a

NL-sentence or a short NL-discourse.

Theme 3: Representation of a query in English as an expression of the query language SPARQL.

The themes of synopses analyzing a selected direction of developing and using natural lan-

guage processing systems

Theme 1: The analysis of the current state of the studies on semantics-oriented computer pro-

cessing of natural language.

Theme 2: The approaches to constructing natural language interfaces for interaction with Linked

Open Data.

Theme 3: The analysis of the current state of the studies on using natural language interfaces for

intelligent consulting.

Theme 4: The approaches to semantic search of information on the Web and on corporative com-

puter nets.

Theme 5: The significance and the approaches to the formation of ontology with the help of nat-

ural language processing systems.

Theme 6: The purpose and significance of designing transformers of the collections of OWL-ex-

pressions into a natural language text.

13

Theme 7: The approaches to constructing and using semantic annotations of electronic docu-

ments.

9. Methods of Instruction

The following educational technologies are used for the realization of various kinds of

teaching/learning processes: lectures, presentations of students’ reports, discussions, solutions to

tasks, provision of additional possibilities of studying theoretical materials for the students want-

ing more deeply learn the discipline.

10. Special Equipment and Software Support

Software

In order to solve practical tasks and to prepare presentations, the students use modern

teaching laboratory basis, including:

standard packets of applied computer programs, in particular:

o informational systems for the preparation of texts (Microsoft Word);

o the systems for the preparation of presentations (Microsoft PowerPoint).

The lecturers use personal computers/notebooks and a head projector for delivering lec-

tures and conducting practical lessons, and also use a hardware of computer classes.

Distance support of the course

The means of distance support provided by the system LMS are employed. The principal

role in distance support is played by theoretical materials for obligatory learning, theoretical ma-

terials for optional studies, the tasks for homed works, and questionnaires,

14