Dr. Dr. rer. nat. Roman Sergienko
Dr. Dr. rer. nat.
Roman
Sergienko
Graduation Date: 03.06.2016
First Employment: riskmethods, Germany
Senior Researcher
Dr. Dr. rer. nat. Roman Sergienko
Theses
Topic
Text Classification for Spoken Dialogue Systems
Status
completed
Description
Spoken Dialogue Systems (SDS) became a new important form of communication between human and machines. SDS can be applied in different areas: call services, smart houses, interactive education etc. A standard SDS consists of the next main stages:
- Speech recognition.
- Text analysis.
- Dialogue management.
- Text generation.
- Speech synthesis.
Text analysis contains semantic and grammatical analysis of user utterances. Before detailed semantic and grammatical analysis of user utterances it would be better to get general information such as conversation topic. Conversation topic identification can be especially useful for multi-domain SDS. Additionally, topic-related models for semantic and grammatical analysis can be designed. Conversation topic categorization can be formulated as a text classification problem. In this case we have predefined topics (classes) and textual information after speech recognition.
From user speech it is possible to extract additional important paralinguistic information which can be helpful for effective human-machine dialogue management:
- User emotion.
- User verbal intelligence.
For these problems different information modalities can be used: acoustic information, video information, and textual information. With multi-modal approaches, fusion of different types of features is performed. User emotion identification, user verbal identification, and human-machine interaction quality estimation based on textual information can be also treated as text classification problems.
Therefore, there are different text classification problems which are important for design of effective and adaptive SDS. The main feature of the considered text classification problems is a fact that textual information is extracted from spoken language.
The main goal of the proposed PhD thesis is searching the most effective approaches and improvement of the existing approaches for text classification problems based on spoken language in the field of spoken dialogue systems.
Automatic Fuzzy Classifiers Forming with Self-tuning Coevolutionary Algorithms
Status
Promotion in Siberian Federal University, Krasnoyarsk, Russia, 10.12.2010
Description
A new method of Michigan and Pittsburgh approaches combining for fuzzy classifier design with evolutionary algorithms is presented. Fuzzy classifier design consists of four stages. The first stage is standard fuzzification. The second one is special procedure of initial rules forming with a priori information from a learning sample. At the third stage Michigan method is used and it provides fast search of fuzzy rules with the best grade of certainty values for different classes and smoothing of randomness at initial population forming. At the fourth stage Pittsburgh method provides rules subset search with the best performance and predefined number of the rules and doesn’t require a lot of computational power. Besides self-tuning cooperative-competitive coevolutionary algorithm for strategy adaptation is used on Michigan and Pittsburgh stages of fuzzy classifier design. This algorithm automatically solves the problem of genetic algorithm parameters setting. Thereby the method allows getting compact fuzzy rule set with appropriate classification performance and with high computation speed. Classification results for machine learning problems from UCI repository and comparison with different alternative classifiers are presented.
Research Interests
- Intellectual Data Analysis
- Intellectual Information Technologies
- Fuzzy Systems
- Evolution Optimization Algorithms
- Machnie Learning
- Knowledge Discovery
- Expert Systems
- Natural Language Processing
- Text Categorization
Projects
1. “Coevolutionary Algorithms Research for Intellectual Information Technologies generating”, Russian Ministry of Education and Science Fund of Extension Work at Scientific and Technical Area, 2008-2011.
2. “Fuzzy Classifiers Collectives Research and Applying for Earth Remote Sensing Images Recognition”, Krasnoyarsk Regional Scientific Fund, 2011.
3. " Feature Extraction and Classification with Fuzzy Logic Technologies for Speech Recognition Problems", DAAD and Russian Ministry of Education amd Science, "Michail Lomonosov Program", Ulm University, 2012-2013.
4. “Development of methods and technologies of intellectual data analysis algorithms parallelization in multiprocessor and distributed computational systems”, Russian Ministry of Education and Science Grant, 2012-2013.
O. Akhtiamov, R. Sergienko and W. Minker
An Approach to off-Talk Detection based on Text Classification within an Automatic Spoken Dialogue System
13th International Conference on Informatics in Control, Automation and Robotics (ICINCO 2016), Lisbon, Portugal, Vol. 2, pp. 288-293, July 2016
Link to Document
Bibtex
A. Spirina, M. Sidorov, R. Sergienko and A. Schmitt
First Experiments on Interaction Quality Modelling for Human-Human Conversation
Proceedings of the 13th International Conference on Informatics in Control, Automation and Robotics (ICINCO 2016), Lisbon, Portugal, Vol. 2, pp. 374-380, July 2016
Bibtex
R. Sergienko, I. Kamshilova, E. Semenkin and A. Schmitt
Weighted Voting of Different Term Weighting Methods for Natural Language Call Routing
Proceedings of the 13th International Conference on Informatics in Control, Automation and Robotics (ICINCO 2016), Lisbon, Portugal, Vol. 1, pp. 38-46, July 2016
Bibtex
R. Sergienko, M. Shan, W. Minker and E. Semenkin
Topic Categorization Based on Collectives of Term Weighting Methods for Natural Language Call Routing
Journal of Siberian Federal University. Mathematics & Physics, Num. 2, pp. 235-245, June 2016
Link to Document
Bibtex
R. Sergienko, M. Shan and W. Minker
A Comparative Study of Text Preprocessing Approaches for Topic Detection of User Utterances
Proceedings of the 10th edition of the Language Resources and Evaluation Conference (LREC 2016), Portorož (Slovenia), pp. 1826-1831, May 2016
Link to Document
Bibtex
A. Spirina, M. Sidorov, R. Sergienko, E. Semenkin and W. Minker
Human-Human Task-oriented Conversations Corpus for Interaction Quality Modeling
Bulletin of Siberian State Aerospace University, Vol. 17, Num. 1, pp. 84-90, March 2016
Link to Document
Bibtex
R. Sergienko, M. Shan and A. Schmitt
A Comparative Study of Text Preprocessing Techniques for Natural Language Call Routing
Proceedings of the 7th International Workshop On Spoken Dialogue Systems (IWSDS), January 2016
Bibtex
R. Sergienko, T. Gasanova, E. Semenkin and W. Minker
Collectives of Term Weighting Methods for Natural Language Call Routing
Informatics in Control, Automation and Robotics, Springer International Publishing Switzerland, Series: Lecture Notes in Electrical Engineering, pp. 99-110, January 2016
Link to Document
Bibtex
R. Sergienko and A. Schmitt
Verbal Intelligence Identification Based on Text Classification
Proceedings of the Annual Conference of the International Speech Communication Association (INTERSPEECH), Dresden, Germany, pp. 2524-2528, September 2015
Link to Document
Bibtex
R. Sergienko, O. Akhtiamov, E. Semenkin and A. Schmitt
A novel approach to neural network design for natural language call routing
12th International Conference on Informatics in Control, Automation and Robotics (ICINCO 2015), Colmar, France, Vol. 01, pp. 102 - 109, July 2015
Link to Document
Bibtex
R. Sergienko, M. Shan, A. Khan, T. Gasanova and W. Minker
Feature selection for text classification based on constraints for term weights
Bulletin of Siberian State Aerospace University, Vol. 16, Num. 1, pp. 119-123, 2015
Link to Document
Bibtex
W. Minker, E. Semenkin, R. Sergienko, R. Lüker and A. Voroshilova
Scientific cooperation between Ulm University and Siberian State Aerospace University
Reshetnev's Readings, Krasnoyarsk, Russian Federation, Vol. 2, pp. 489-491, November 2014
Link to Document
Bibtex
S. Akhmedova, E. Semenkin and R. Sergienko
Automatically Generated Classifiers for Opinion Mining with Different Term Weighting Schemes
Proceedings of the 11th International Conference on Informatics in Control, Automation and Robotics - ICINCO 2014, Vienna, Austria, Vol. 2, pp. 845-850, September 2014
Link to Document
Bibtex
T. Gasanova, R. Sergienko, E. Semenkin and W. Minker
Dimension Reduction with Coevolutionary Genetic Algorithm for Text Classification
Proceedings of the 11th International Conference on Informatics in Control, Automation and Robotics (ICINCO), Vienna University of Technology, Austria, Vol. 1, pp. 215-222, September 2014
Link to Document
Bibtex
R. Sergienko, T. Gasanova, E. Semenkin and W. Minker
Text Categorization Methods Application for Natural Language Call Routing
Proceedings of the 11th International Conference on Informatics in Control, Automation and Robotics - ICINCO 2014, Vienna, Austria, Vol. 2, pp. 827-831, September 2014
Link to Document
Bibtex
T. Gasanova, R. Sergienko, S. Akhmedova, E. Semenkin and W. Minker
Opinion Mining and Topic Categorization with Novel Term Weighting
Proceedings of the 5th Workshop on Computational Approaches to Subjectivity, Sentiment and Social Media Analysis, Association for Computational Linguistics, Baltimore, Maryland, USA., pp. 84–89, June 2014
Link to Document
Bibtex
T. Gasanova, R. Sergienko, W. Minker, E. Semenkin and E. Zhukov
A Semi-supervised Approach for Natural Language Call Routing
Proceedings of the SIGDIAL 2013 Conference, pp. 344-348, August 2013
Link to Document
Bibtex
R. Sergienko and E. Semenkin
Multistep fuzzy classifier design with self-tuning coevolutionary algorithm
Proceedings of the 10th International Conference in Informatics in Control, Automation and Robotics (ICINCO 2013), Vol. 1, pp. 113-120, July 2013
Link to Document
Bibtex
R. Sergienko, W. Minker, E. Semenkin and T. Gasanova
Call Routing Problem Solving Method Based on a New Term Relevance Estimation
Program Products and Systems, Num. 1, pp. 90-93, March 2013
Link to Document
Bibtex
T. Gasanova, R. Sergienko, W. Minker and E. Zhukov
A New Method for Natural Language Call Routing Problem Solving
Bulletin of Siberian State Aerospace University named after academician M.F. Reshetnev, Num. 4, 2013
Link to Document
Bibtex
R. Sergienko and E. Semenkin
Michigan and Pittsburgh methods combination for fuzzy classifier design with coevolutionary algorithm
Proc. of 2013 IEEE Congress on Evolutionary Computation, Cancun, Mexico, June 2013, pp. 3252-3259, 2013
Link to Document
Bibtex
T. Gasanova, R. Sergienko, W. Minker and E. Semenkin
Text Categorization by Coevolutionary Genetic Algorithm
Bulletin of Siberian State Aerospace University named after academician M.F. Reshetnev, Num. 4, pp. 104-108, 2013
Link to Document
Bibtex
R. Sergienko
Fuzzy Classifier Design with Coevolutionary Algorithms Applying for Speaker Identification
Bulletin of Siberian State Aerospace University named after academician M.F. Reshetnev, Num. 4, pp. 98-104, 2012
Link to Document
Bibtex
R. Sergienko and E. Semenkin
Multistep Fuzzy Classifier Forming with Cooperative-Competitive Coevolutionary Algorithm
Advances in Swarm Intelligence. Third International Conference, ICSI 2012, Shenzhen, China, June 17-20, 2012 Proceedings, Part I, Springer Berlin Heidelberg, Series: Lecture Notes in Computer Science, Vol. 7331, 2012
Link to Document
DOI
Bibtex
R. Sergienko, E. Semenkin and V. Bukhtoyarov
Evolutionary Approach for Automatic Design of Neural Networks Ensembles for Modeling and Time Series Forecasting
Proceedings of the IADIS International Conference Intelligent Systems and Agents 2011, pp. 93-96, 2011
Link to Document
Bibtex
R. Sergienko, E. Semenkin and V. Bukhtoyarov
Hybrid Fuzzy Classifier Design with Coevolutionary Genetic Algorithm
Proceedings of the IADIS International Conferences Informatics 2011, pp. 35-42, 2011
Link to Document
Bibtex
R. Sergienko and E. Semenkin
Competitive Cooperation for Strategy Adaptation in Coevolutionary Genetic Algorithm for Constrained Optimization
Proc. of 2010 IEEE Congress on Evolutionary Computation, pp. 1626-1631, 2010
Link to Document
DOI
Bibtex
R. Sergienko
Fuzzy classifier forming method with self-tuning coevolutionary algorithms
Artificial intelligence and decision-making, Num. 3, pp. 98-106, 2010
Link to Document
Bibtex
S. Ultes, T. Gasanova, R. Sergienko, M. Sidorov, D. Zaykovskiy and W. Minker
Research at Ulm University
Presentations at Siberian Federal University and Siberian State Aerospace University, Krasnoyarsk (Russia), September 2013
Lectures
H. Hofmann, K. Jokinen, F. Nothdurft, T. Gasanova, M. Sidorov, R. Sergienko, S. Ultes and W. Minker
Selected Topics In Spoken Dialogue Research at Ulm University and Tartu University
Graduate Course within the ERASMUS/SOCRATES Mobility Programme, Institute of Communications Engineering, University of Ulm, November 2013
Open Topics
For open topics for Diploma, Master or Bachelor Theses and Studienarbeiten, please come and see me in my office or write an e-mail.
In Progress / Completed
Master Thesis
Optimization of weights for term clustering in text classification (Euler Program)
2016
Link to Document
Master Thesis
Optimization of weights for voting in term weighting collectives for text classification (Euler Programme)
2016
Link to Document
Master Thesis
Feature Selection and Feature Transformation for Natural Language Call Routing
2015
Master Thesis
Fuzzy classification methods application for hierarchical conversation topic categorization (Euler Program)
2015
Link to Document
Master Thesis
Neural networks application for hierarchical conversation topic categorization
2015
Link to Document