Special Sessions

  • S1. Theoretical and Applied Aspects of Imprecise Probabilities
  • S2. Fuzzy Implication Functions
  • S3. Data Perspectivism in Ground Truthing and Artificial Intelligence
  • S4. Fuzzy Methods in Data Mining and Knowledge Discovery
  • S5. Decision Making Modeling and Applications
  • S6. Aggregation Theory beyond the Unit Interval
  • S7. Soft Methods in Statistics and Data Analysis
  • S8. Logical Structures of Opposition and Logical Syllogisms
  • S9. Mathematical Fuzzy Logics: Modalities, Quantifiers and Uncertainty
  • S10. Machine Learning for partially-labeled Data
  • S11. Information Fusion Techniques based on Aggregation Functions, Pre-aggregation Functions, and their Generalizations
  • S12. Managing Imprecise Information for XAI
  • S13. Uncertainty in Medicine
  • S14. Fuzzy Mathematical Analysis and its Applications
  • S15. Uncertainty, Fuzziness and Many-valued Logic
  • S16. Fuzziness in Smart Cities
  • S17. Natural Language and Information Systems
  • S18. Interval Uncertainty
  • S19. Uncertainty, Reliability and Explainability in AI
  • S20. Soft Computing and Artificial Intelligence Techniques in Image Processing, with the focus on Advanced Shape Analysis and Reconstruction

s1. theoretical and applied aspects of Imprecise Probabilities

Organizers: Enrique Miranda and Ignacio Montes (both the organizers are with the University of Oviedo, Spain)

Description: This session is devoted to Imprecise Probability Theory. This theory involves all the mathematical models that can be used as more flexible tools than usual Probability Theory when the available information is scarce, vague, or incomplete. It includes lower previsions, n-monotone capacities, belief functions, possibility measures, or non-additive measures, among others.

This special session aims to include papers related to Imprecise Probabilities that either present a significant advance in the foundations or show potential applications in real problems. In addition, papers, where the connection between imprecise probability theories and other fields such as fuzzy sets or game theory is emphasized, are also welcome.


s2. Fuzzy Implication Functions

Organizers: Michal Baczynski (University of Silesia in Katowice, Poland), Balasubramaniam Jayaram (Indian Institute of Technology Hyderabad, India), and Sebastia Massanet (University of the Balearic Islands, Spain)

Description: For more than a decade now, fuzzy implication functions have become one of the main research lines of the fuzzy logic community. These logical connectives are the generalization of the classical two-valued implication to the infinite-valued setting. In addition to modeling fuzzy conditionals, they are also used to perform backward and forward inferences in different fuzzy rule-based systems. Moreover, they have proved to be useful not only in fuzzy control and approximate reasoning but also in many other fields such as Multi-Valued Logic, Image Processing, Data Mining, Computing with Words, and Rough Sets, among others.

Due to this great variety of applications, fuzzy implication functions have attracted the efforts of many researchers from the points of view of both theory and applications. Indeed, the theoretical perspective focuses on problems whose solutions provide important insights from the point of view of their applications. Therefore, this special session seeks to bring together researchers interested in recent advances in the theory and the applications of fuzzy implication functions, concerning, among others, characterizations, representations, generalizations, and their relationships with fuzzy negations, triangular norms, uninorms and other fuzzy logic connectives.


s3. Data Perspectivism in Ground Truthing and Artificial Intelligence

Organizers: Andrea Campagner (University of Milano-Bicocca, Italy), Teresa Scantamburlo (European Centre for Living Technology, Ca’ Foscari University of Venice, Italy), Valerio Basile (University of Turin, Italy), and Federico Cabitza (University of Milano-Bicocca, Italy)

Description: Many Artificial Intelligence applications are based on supervised machine learning (ML), which ultimately grounds on manually annotated data. The annotation process (often called
ground-truthing) is often performed in terms of a majority vote and this has been proved to be often problematic, as highlighted by recent studies on the evaluation of ML models. Recently, a different paradigm for ground-truthing has started to emerge, called data perspectivism [1], which moves away from traditional majority aggregated datasets, towards the adoption of methods that integrate different opinions and perspectives within the knowledge representation, training, and evaluation steps of ML processes, by adopting a non-aggregation policy. This alternative paradigm obviously implies a radical change in how we develop and evaluate ML systems: such ML systems have to take into account multiple, uncertain, and potentially mutually conflicting views [2]. This obviously brings both opportunities and difficulties: novel models or training techniques may need to be designed, and the validation phase may become more complex. Nonetheless, initial works have shown that data perspectivism can lead to better performances [3,4], and could also have important implications in terms of human-in-the-loop and interpretable AI, as well as in regard to the ethical issues or concerns related to the use of AI systems [5]. Data perspectivism is a framework to treat uncertainty (the main theme of IPMU) at the level of knowledge modeling and its integration in the development and evaluation of systems.

The scope of this special session is to attract contributions related to the management of subjective, uncertain, multi-perspective, or otherwise non-aggregated data in ground-truthing, machine learning, and more generally artificial intelligence systems.

Topics of interest include:

  • Subjective, uncertain, or conflicting information in annotation and crowdsourcing processes;
  • Limits and problems with standard data annotation and aggregation processes;
  • Theoretical studies on the problem of learning from multi-rater and non-aggregated data;
  • Participation mechanisms/incentives/gamification for rater engagement and crowdsourcing;
  • Ethical and legal concerns related to annotation and aggregation processes in ground-truthing;
  • Creation and documentation of multi-rater and non-aggregated datasets and benchmarks;
  • Development of ML algorithms for multi-rater and non-aggregated data;
  • Development of techniques to detect and manage multiple forms of uncertainty in multi-rater and non-aggregated data;
  • Techniques for the evaluation of ML systems based on multi-rater and non-aggregated data;
  • Applications of data perspectivism and non-aggregated data to interpretable, human-in-the-loop AI and algorithmic fairness;
  • Experimental and application studies of ML/AI systems on multi-rater and non-aggregated data, in possibly different application domains (e.g. NLP, medicine,legal studies, etc.)

We believe that this special session could be of interest to researchers and practitioners interested in: machine learning and artificial intelligence; uncertainty representation and management (including: fuzzy set theory, belief functions, and imprecise probabilities); crowdsourcing; interpretable, explainable, and human-in-the-loop AI; ethical, philosophical and legal aspects of AI and ML; social choice theory; aggregation operators.


s4. Fuzzy methods in Data Mining and Knowledge Discovery

Organizers: Carlos J. Fernández-Basso, Karel Gutiérrez Batista, and José Ángel Diaz-Garcia (all the organizers are with the University of Granada, Spain)

Description: The objective of the special session is to provide a forum for the discussion of recent advances in the application of Data Mining and Knowledge Discovery technologies to diverse problems, focusing on those involving fuzzy methods, and to offer an opportunity for researchers to identify new and promising research directions. Data Mining aims at the automatic discovery of underlying non-trivial knowledge from datasets by applying intelligent analysis techniques. The interest in this research area has experienced considerable growth in the last years due to two key factors: (a) knowledge hidden in organizations’ databases can be exploited to improve strategic and managerial decision-making in the current ultra-competitive markets; (b) the large volume of data managed by organizations makes it impossible to carry out an analysis process manually.

Nowadays, the volume of information digitally stored has considerably increased not only in database format but also in text format which is available in open source bases such as the Web, including log files registering the use of the information or social media content. This has contributed to increasing the interest in Text and Web Mining techniques. On one hand, these techniques aim to automatize the analysis process by introducing a variety of intelligent techniques to learn, optimize and represent uncertain and imprecise knowledge. On the other hand, these tools offer the possibility to analyze massive data offering more efficient algorithms and a suitable selection of obtained results in terms of their novelty, usefulness, and interoperability.

Topics of interest include, but are not limited to, the following topics:

  • Data, text, and Web mining
  • Stream data mining
  • Temporal data series
  • Big data mining
  • Imprecision, uncertainty, and vagueness in data mining
  • Data pre- and post-processing in data mining
  • Parallel and distributed data mining algorithms
  • Information summarization and visualization
  • Human-machine interaction for data access
  • Semantic models to represent input data and extracted knowledge in a Data Mining process
  • Applications of Data Mining techniques: health, tourism, biological process, customer profiles, anomaly detection, emergency management, situation recognition, etc.

S5. Decision making modeling and applications

Organizers: Rocío de Andrés Calle and Silvia Prieto Herraez (both the organizers are with the University of Salamanca, Spain)

Description: This session is dedicated to the discussion of recent advances in Decision Making. The cross-cutting nature of this research area has led to the development of different models from different approaches and perspectives in a wide range of disciplines such as Psychology, Economics, Political Sciences, Social Choice, Operations Research, Medicine, Artificial Intelligence, Engineering, etc. This special session aims at providing an opportunity for researchers working in this research area to discuss and disseminate theoretical and empirical advances in fundamental, approaches, methodologies, software systems, and applications, to share their novel ideas, original research results, and practical experiences.

Topics of interest include but are not limited to:

  • Decision analytic models and applications
  • Decision probabilistic models and applications
  • Analyzing and presenting simulation output from Decision modeling
  • Decision modeling under uncertainty and applications
  • Technology for Decision modeling and applications
  • Other related topics

S6. Aggregation theory beyond the unit interval

Organizers: Bernard De Baets (Ghent University, Belgium) and Raúl Pérez-Fernández (University of Oviedo, Spain)

Description: The study of means has been of interest to mathematicians since ancient times and its formalization is normally attributed to Cauchy in 1821. This interest in means and, more generally, in aggregation processes on the unit interval experienced a boost of attention with the foundation of fuzzy set theory, resulting in the development of the subfield of aggregation theory. Since then, with the increasing availability of data, aggregation theorists have become interested in the aggregation on more general structures than the unit interval, such as partially ordered sets and lattices, and even some application fields have faced the need of aggregating data of a very different nature. Prototypical examples are the fields of computer science (aggregation of strings), computer vision (aggregation of images), decision-making (aggregation of fuzzy relations), geochemistry (aggregation of compositional data), multivariate statistics (aggregation of vectors), and social choice theory (aggregation of rankings).

There is little doubt that aggregation is a core topic of the IPMU conference, as is proven by the fact that many special sessions on the topic have taken place at past editions of the conference. The aim of this IPMU 2022 Special Session is to continue with this longstanding tradition and to offer researchers in aggregation theory the opportunity to keep up to date with the most recent advances in the field. All contributions dealing with the aggregation of different structures, from either a theoretical, practical, and algorithmic point of view, are welcome.


S7. Soft Methods in Statistics and Data Analysis

Organizers: Przemyslaw Grzegorzewski and Katarzyna Kaczmarek-Majer (both the organizers are with the Systems Research Institute, Polish Academy of Sciences, Poland)

Description: A growing diversity of data and the increasing complexity of systems developed in science and engineering reveal the need for more flexible tools to handle uncertainty. It becomes more and more clear that challenges faced by data analysts require soft modeling and soft computing methods that are less rigid than traditional approaches and therefore which adapt more easily to the actual nature of the available data. The desired methodology should combine different types and aspects of uncertainty, including randomness, imprecision, ambiguity, etc. By integrating fuzzy logic, probability theory, and other approaches, more robust and interpretable models and tools can be developed that better capture all kinds of information contained in the data.

The aim of this Special Session is to bring together theorists and practitioners who apply soft methods in statistical reasoning and data analysis to exchange ideas and discuss new trends that enrich traditional approaches and tools.

Topics of interest include but are not limited to:

  • Analysis of censored or missing data
  • Analysis of fuzzy data
  • Bayesian methods
  • Clustering and classification
  • Data mining
  • Fuzzy random variables
  • Fuzzy regression methods
  • Granular computing
  • Imprecise probabilities
  • Interval data
  • Machine learning
  • Possibility theory
  • Random sets
  • Robust statistics
  • Semi-supervised learning
  • Soft computing
  • Statistical software for imprecise data
  • Time series analysis and forecasting

S8. Logical structures of opposition and logical syllogisms

Organizers: Vilém Novák (University of Ostrava, Czech Republic), Petra Murinová (University of Ostrava, Czech Republic), and Stefania Boffa (University of Milano-Bicocca, Italy)

Description: Human reasoning and logical derivation is a core area in a wide range of disciplines such as Artificial Intelligence, Engineering, Psychology, Economics, Political Sciences, Medicine, etc. The objective of this special session is to provide a forum for the discussion of recent advances in the application of the theory of syllogistic reasoning and the corresponding logical structures of opposition. There are several logical characterizations of different structures of opposition that extend Aristotle’s square of opposition. This special session aims at providing an opportunity for researchers working in this research area to discuss fundamental approaches, methodologies, software systems, and applications, and to share their novel ideas, original research results, and practical experiences.

We invite contributions that extend traditional structures of opposition and possible applications of these structures. We also welcome contributions regarding applications of the theory of syllogistic reasoning:

  • The theory of generalized quantifiers and their special cases (e.g., fuzzy or intermediate ones)
  • Logical structures of opposition
  • The theory of syllogistic reasoning with generalized and fuzzy quantifiers
  • Mining knowledge from data in the form of special rules (e.g., association or fuzzy/linguistic ones)
  • Mining knowledge from time series
  • Concept analysis with the help of fuzzy quantifiers

S9. Mathematical Fuzzy Logics: Modalities, Quantifiers, and Uncertainty

Organizers: Tommaso Flaminio, Lluis Godo, Sara Ugolini, and Amanda Vidal (all the organizers are with the Artificial Intelligence Research Institute IIIA – CSIC, Spain)

Description: The study of logical, algebraic, proof-theoretic tools for the management of modes of truth, quantifiers, and uncertainty is a well-established line of research, whose development has influenced significantly many areas of pure and applied research. This special session titled Mathematical Fuzzy Logics: Modalities, Quantifiers, and Uncertainty, aims at collecting papers connected to modal expansions, first and higher-order, and uncertainty theories based on the formal setting of Mathematical Fuzzy Logic.

In many real-world situations, we need to evaluate the degree of uncertainty, or to estimate the feasibility, of a sentence (or a class of sentences) that cannot be exactly regarded as completely true or false, without a sensible lack of precision. Many-valued logics, fuzzy logics, and their algebraic semantics offer the formal tools to deal with graded truth and imprecise events. Modalities, quantifiers, and theories of uncertainty that are built over those logico-algebraic structures that model many valued events are adequate tools to model those situations in which a quantitative evaluation of the combination of uncertainty and imprecision is needed.

This special session will focus on (but will not be limited to) the following topics:

  • Modal fuzzy logic
  • First order fuzzy logics
  • Generalized quantifiers
  • States and internal states on algebras of fuzzy logics
  • Generalized uncertainty measures on fuzzy events
  • Philosophical foundations of uncertainty and fuzziness

S10. Machine Learning for Partially Labeled Data

Organizers: Andrea Campagner (University of Milano-Bicocca, Italy), Nahuel Costa (University of Oviedo, Spain), and Luciano Sánchez (University of Oviedo, Spain)

Description: Machine learning from incomplete data makes it possible to use imperfect and low-cost data to create a robust predictive model. This type of learning is oriented towards uncertain data, and as such is related to different themes of the IPMU’2022 conference. These include weak supervised learning and its variants, such as incomplete supervision, where only a subset of training data is labeled; inexact supervision, where the training data is coarse-grained labeled; and inaccurate supervision, where the labels are not always true.

Other related fields are active learning, which measures the information that new data would bring to the model based on its imprecision and the cost of obtaining it, or semi-supervised learning, which uses assumptions about the structure of imprecise data to exploit unlabelled data. Contrastive learning and many other different techniques can also benefit from the application of machine learning to partially labeled data.

This special session is intended to serve as a common space for researchers in this field to share their latest findings on incompletely labeled data, including the development of practical algorithms to address different tasks based on such kind of data, as well as the conceptual foundations of this area.


S11. Information Fusion Techniques Based on Aggregation Functions, Preaggregation Functions and their Generalizations

Organizers: Humberto Bustince (Universidad Publica de Navarra – UPNA, Spain), Graçaliz P. Dimuro (Universidade Federal do Rio Grande, Brazil), Javier Fernández (Universidad Publica de Navarra – UPNA, Spain), and Tiago da Cruz Asmus (Universidade Federal do Rio Grande, Brazil)

Description: The search for new information fusion techniques is currently a hot topic in almost every research field, from image processing and decision making to deep learning. This interest has led to a new analysis of the notion of aggregation function, as well as to the introduction of new concepts that go beyond usual aggregation functions, either by considering more general definitions (e.g., considering weaker forms of monotonicity, as directional monotonicity and ordered directional monotonicity), or by extending them to other frameworks different from that of the unit interval (e.g., intervals, lattices, any closed real interval).

The study of such generalizations has led to more flexibility in applications, from the control of the uncertainty to novel forms of fusioning data. The ongoing study of these concepts as well as of their applications has already been displayed in many special sessions in previous IPMU conferences, as well as other conferences such as EUSFLAT and FUZZ-IEEE, where we verified an expressive number of participants and the attendance of the main researchers of the area. Several past special sessions gave birth to special issues of prestigious journals, edited by the organizers.

So, the aim of this IPMU special session is to follow this longstanding tradition and to present a forum for researchers to discuss the most up-to-date research in the field of fusion techniques using aggregation functions, preaggregation functions, and their generalizations/extensions, as well as their possible applications in any field of artificial intelligence and computer science, including, but not limited to, image processing, classification, recurrent neural networks, deep learning, big data, approximate reasoning, computational brain or decision-making.


s12. MIIXAI: Managing Imprecise Information for XAI

Organizers: Dario Malchiodi (University of Milan, Italy) and Corrado Mencar (University of Bari “Aldo Moro”, Italy)

Description: People have an exceptional ability in managing imprecise information in forms that are well captured by several theories within the Granular Computing paradigm, such as Fuzzy Set Theory, Rough Set Theory, Interval Computing, and hybrid theories among others. Endowing XAI systems with the ability to deal with the many forms of imprecision, is therefore a key challenge that can push forward current XAI technologies towards more trustworthy systems based on imprecise information (II) and full collaborative intelligence.

The Special Session will gather recent advancements in topics like foundational, theoretical and methodological aspects of imprecision management in XAI, new technologies for representing and processing imprecision in XAI systems, as well as real-world applications that demonstrate explainability improvements through imprecision management.

The covered topics include, but are not limited to (alphabetically ordered):

  • Design of explainable II-based systems
  • Evaluation of explainability in models for II
  • Hybrid systems dealing with different forms of imprecision
  • Successful applications of explainable II-based systems
  • Induction of explainable models from II
  • Theoretical aspects of explainability in II-based systems

S13. Uncertainty in Medicine

Organizers: Krzysztof Dyczkowski (Adam Mickiewicz University in Poznań, Poland), Barbara Pękala (University of Rzeszów, Poland), and Anna Wilbik (Maastricht University, The Netherlands)

Description: Uncertainty is a serious problem in everyday medical practice and it is observed and described in the medical literature. There are many meanings and types of imprecision in medicine with each of them having a different effect on the diagnosis or treatment. The imprecision types may be divided into objective (caused by the complexity or nature of the phenomenon) or subjective (caused by a personal opinion or doctor’s interpretation) or caused by the low quality of the information, e.g., due to incompleteness of the data. Through the use of soft computing methods, it is possible to construct models that, when uncertainty is taken into account properly, increase the effectiveness of medical decision-making.

The session aims at exchanging the experiences of scientists struggling with uncertainty in medicine and discussing possible future developments in this area.

Potential topics of interest include but are not limited to:

  • Intelligent decision-making systems in medical environments
  • Machine learning based on medical data
  • Uncertain data aggregation and fusion
  • Applications of Fuzzy sets, Rough sets, and their extensions
  • Modeling different types of uncertainty in medical data
  • Assessment of medical data and models quality

S14. Fuzzy Mathematical Analysis and its Applications

Organizers: Yurilev Chalco-Cano (Universidad de Tarapacá, Chile), Beatriz Hernández-Jiménez (University Pablo de Olavide, Spain), and Rosana Rodríguez-López (Universidade de Santiago de Compostela, Spain)

Description: Mathematical analysis, in the broad sense of the term, includes a very large part of mathematics. It includes the theory of real variable functions, differential calculus, integral calculus, approximation theory, the theory of ordinary and partial differential equations, or the theory of integral equations, among others. On the other hand, fuzzy mathematics forms a branch of mathematics related to fuzzy set theory and fuzzy logic. Fuzzy mathematical analysis appears as one of the natural ways to handle the uncertainty in related concepts of mathematical analysis. In the recent decades, the concepts of fuzzy metric space, fuzzy integral and derivatives, or fuzzy initial value problems have been proposed.

The goal of this session is to bring together researchers interested in recent advances in fuzzy mathematical analysis and its applications.

The topics of this special session include, but are not limited to, the following:

  • Spaces of fuzzy sets
  • Fuzzy arithmetic
  • Fuzzy metric spaces
  • General fuzzy-valued functions
  • Convexity and fuzzy-valued functions
  • Differentiability of fuzzy-valued functions
  • Fuzzy integrals
  • Fuzzy differential equations
  • Fuzzy partial differential equations
  • Fuzzy optimal control
  • Fuzzy systems
  • Fuzzy optimization problems

S15. Uncertainty, fuzziness and many-valued logic

Organizers: Stefano Aguzzoli (University of Milan, Italy), Matteo Bianchi (University of Milan, Italy), and Brunella Gerla (University of Insubria, Italy)

Description: Many-valued logics have constituted for several decades key conceptual tools for the formal description and management of fuzzy, vague and uncertain information. In the last few years, the study of these logical systems has seen a bloom of new research related to the most diverse areas of mathematics and applied sciences. Relevant recent developments in this field are connected to the natural semantics of non-classical events. The combination in a unique conceptual framework of the logic and the probability of a class of non-classical events, usually reached through the algebraic semantics and their topological or combinatorial dualities, provides both the theoreticians and the application oriented scholars with powerful tools to deal with this kind of events.

This special session is devoted to the most recent development in the realm of many-valued logics, with particular emphasis on theoretical advances related to algebraic or alternative semantics, combinatorial aspects, topological and categorical methods, proof theory and game theory, many-valued computation. In particular, results directed towards a better understanding of the natural semantics of non-classical events will be appreciated. Further, a special attention is also given to connections and synergies between many-valued logics and other different formal approaches to vague and approximate reasoning, such as Rough Sets, Formal Concept Analysis and Relational Methods.

A partial list of topics is the following:

  • Algebraic semantics of many-valued logics
  • Applications of many-valued logics to Formal Concept Analysis and Relational Methods
  • Applications of many-valued logics to Fuzzy Sets and to Rough Sets
  • Combinatorial or topological dualities
  • Computational complexity of many-valued logics
  • Many-valued computational models
  • Modal logic approaches to probability and uncertainty in many-valued logics
  • Natural and alternative semantics for many-valued logics
  • Proof theory for many-valued logics
  • Representation theory
  • Subjective probability approaches to many-valued logics and nonclassical events

S16. Fuzziness in Smart Cities

Organizers: Allel Hadjali (National Engineering School for Mechanics and Aerotechnics, France), Miroslav Hudec (Technical University of Ostrava, Czech Republic / University of Economics in Bratislava, Slovakia), and Edy Portmann (University of Fribourg, Swisse)

Description: The activities in smart cities are focused mainly on effectiveness. Although remarkable, it is not sufficient. We should also consider citizens and their needs. Fuzzy logic and soft computing solutions should allow smart participation (surveys, voting, managing fuzzy data), data analyses (classification, clustering …), data mining from data in an explainable and understandable way for all urban stakeholders (by, e.g., linguistic summarizations), and predicting future impacts (e.g., new constructions on environmental burden). Fuzzy logic brings, in general, explainable solutions, but the interpretability of complex models should be evaluated. This results in services that are based on urban stakeholders needs and thus sustainably improve the quality of coexistence in a city.

Relevant topics include the following (but are not limited to):

  • Explainable classification by fuzzy systems
  • Fuzzy logic in data collection
  • Linguistically summarizing and interpreting data
  • Predictive and advanced machine learning models
  • Modelling environmental developments
  • Smart participation of citizens
  • Smart governance supported by fuzzy logic and related topics
  • From smart to cognitive cities by fuzzy logic
  • Future of fuzzy logic in smart cities

s17. Natural Language and Information Systems

Organizers: Juan-Luis Castro, Nicolás Marín, and Daniel Sánchez (all the organizers are with the University of Granada, Spain)

Description: Research on Information Systems on the one hand and Natural Language on the other appear regularly in the contributions to the IPMU conferences, particularly since there are many aspects related to the management of uncertainty in both areas. Regarding Information Systems, topics include among others fuzzy databases, flexible querying, and information retrieval. On its turn, Natural Language has appeared through topics like Computing with Words and Perceptions, Knowledge Representation, Linguistic Summarization, generation of interpretable models in Data Mining and Machine Learning, etc.

In addition to contributions in each of these particular areas, this special session is interested in the application of Natural Language for different purposes in Information Systems. The importance of the use of Natural Language in Information Systems can be seen also in the fact that a conference series exists on the topic within the Natural Language Processing community, with 26 editions up to 2021 covering aspects like the application of Machine Learning techniques, Semantic Analysis, Opinion Mining and Sentiment Analysis, Social Networks and Media Intelligent Analysis, Question Answering and Answer Generation, Information Retrieval, Conversational Agents, Natural Language Data Mining and Knowledge Discovery, Ontology Engineering, etc.

In recent years, the potential synergies between Natural Language Processing and the Soft Computing research areas have been explored by members of both scientific communities through special sessions and workshops (see for instance the ECAI 2020 IntelLang Workshop) and research networks involving researchers from both NLP and Soft Computing. In this context, one of the objectives of this special session is to provide a forum for the discussion of research on Natural Language in Information Systems involving the management of uncertainty and other applications of soft computing techniques, in topics like:

  • Fuzzy databases
  • Flexible querying
  • Information retrieval
  • Computing with Words and Perceptions in Information Systems
  • Natural Language for Information Representation
  • Linguistic Summarization of data
  • Interpretability in Data Mining and Machine Learning
  • Opinion Mining and Sentiment Analysis
  • Social Networks and Media Intelligent Analysis
  • Question Answering and Answer Generation in Information Systems
  • Conversational Agents in Information Systems
  • Natural Language-based Data Mining and Knowledge Discovery

s18. Interval Uncertainty

Organizers: Martine Ceberio and Vladik Kreinovich (both the organizers are with the University of Texas at El Paso, USA)

Description: Interval uncertainty is closely related to fuzzy techniques; indeed, if we want to know how the fuzzy uncertainty of the inputs propagates through the data processing algorithm, then the usual Zadeh’s extension principle is equivalent to processing alpha-cuts (intervals) for each level alpha.

This relation between intervals and fuzzy computations is well known, but often, fuzzy researchers are unaware of the latest most efficient interval techniques and thus use outdated less efficient methods. One of the objectives of the proposed session is to help fuzzy community by explaining the latest interval techniques and to help interval community to better understand the related interval computation problems.

Yet another relation between interval and fuzzy techniques is that the traditional fuzzy techniques implicitly assume that experts can describe their degree of certainty in different statements by an exact number. In reality, it is more reasonable to expect experts to provide only a range (interval) of possible values – leading to interval-valued fuzzy techniques that, in effect, combine both types of uncertainty.


s19. Uncertainty, Reliability and Explainability in AI

Organizers: Salem Benferhat and Karim Tabia (both the organizers are with the Artois University, France)

Description: The last decade has been marked noticeably by the growing use of innovative intelligent systems and applications that rely heavily on AI models and implementations. We are then confronted with new problems and risks, in particular the complexity of the systems and their opacity and the sensitivity of certain critical applications. These are mainly issues of confidence, trust, interpretability and explainability of reasoning and decisions that can be taken automatically.

This special session aims to bring together the AI community addressing issues of reliability, uncertainty, interpretability and explainability. In particular, we seek contributions in uncertainty related knowledge representation and reasoning, reliable and explainable AI and applications. This will foster cross fertilization among knowledge representation and reasoning and machine learning. Depending on the number of accepted submissions in this session, a special issue of a journal may be scheduled later.

The topics of interest include (but are not limited to):

  • Uncertain information representation, reasoning under uncertainty, uncertain data aggregation and fusion
  • Decision making under uncertainty
  • Management of imperfect and heterogeneous data
  • Reliability and trust, calibration, uncertainty management in machine learning
  • Links between knowledge representation, reasoning , decision making and machine learning
  • Explainable and interpretable AI
  • Relationships between Knowledge representations/formalisms, Reliability and Explainability
  • Applications of Uncertain, Reliable and Explainable AI Systems

S20. Soft Computing and Artificial Intelligence Techniques in Image Processing, with the focus on Advanced Shape Analysis and Reconstruction

Organizers: Irina Perfilieva (University of Ostrava, Czech Republic), Javier Montero (Universidad Complutense de Madrid, Spain), Humberto Bustince (Universidad Publica de Navarra, Spain), Carlos Lopez-Molina (Universidad Publica de Navarra, Spain), and Jan Hůla (University of Ostrava, Czech Republic)

Description: The Special Session is focused on theoretical and application aspects of image processing.

Traditionally, our interest is in the direction of modern AI methods that use and combine conventional, fuzzy, and neural network methods together.

This year we propose to discuss in detail the topic of shape analysis and reconstruction, which plays a key role in applications ranging from traditional geometry processing to advanced 3D content management techniques. Shape analysis is closely related to semantic understanding, which is extremely important in applications dealing with a 3D representation of a scene.

Among other things, we are looking for tools that support the semantic annotation of digital 3D shapes.

Our interest also lies in the development of shape analysis tools capable of capturing a diverse set of morphologically significant functions, possibly at different scales, for the purpose of segmentation, classification, clustering, and other tasks.

This special session will focus on (but not limited to) the following topics:

  • Soft computing methods in all aspects of image analysis and processing
  • Medical image processing
  • Reliable feature extraction
  • HD Manifolds in image representation
  • Manifold learning
  • Laplace-Beltrami operators and Graph Laplacians
  • Reduction of dimensionality
  • Manifolds and fuzzy sets