List of accepted sessions

SS1 - Decision Making Under Uncertainty and Fuzzy Optimization: Theory, Algorithms, and Applications

Assigned Area: Area 9 - Fuzzy systems: decision-making, inference systems, preference modeling, optimization

ORGANIZERS

• Weldon A. Lodwick (University of Colorado Denver, USA)
• Masahiro Inuiguchi (Osaka University, Japan)

SCOPE

The special session seeks to bring together researchers and talks in the areas of decision making and optimization in which uncertainty, fuzzy, and/or possibility analysis is an inherent part of talk. In particular, this special session, invites talks dealing with decision making, fuzzy/interval/uncertainty decision making, interval, uncertainty, fuzzy, and possibilistic optimization. Talks with theoretical content, algorithms, state-of–the–art surveys, and applications in the above-mentioned areas, are welcome.

The topics associated with session include at least the following areas: decision making, fuzzy decision making, decision making under uncertainty, Analytic Hierarchy Processes (AHP), Interval AHP, Fuzzy AHP, optimization under uncertainty, interval optimization, fuzzy optimization, possibilistic optimization.

SS2 - Interval Uncertainty

Assigned Area: Area 1 - Basics of fuzzy set theory: fuzzy relations, generalizations of fuzzy sets, etc.

ORGANIZERS

• Martine Ceberio (University of Texas at El Paso, USA)
• Vladik Kreinovich (University of Texas at El Paso, USA)

SCOPE

Interval uncertainty is closely related to fuzzy: e.g., Zadeh's extension principle is equivalent to applying interval analysis to alpha-cuts (intervals).

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is well known, but often, fuzzy researchers are unaware of the latest most efficient interval techniques and thus use outdated less efficient methods.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 the fuzzy community by explaining the latest interval techniques and to help the interval community to better understand the related interval computation problems. Another relation is interval-valued fuzzy techniques that combine both uncertainties.

SS3 - Fuzzy Implication Functions

Assigned Area: AGOP - Summer School on Aggregation Operators

ORGANIZERS

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

SCOPE

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 twoag-valued implication to the infinite-valued setting. In addition to modelling 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.

SS4 - Soft Computing for Evolving Data Streams: Advances in Real-Time Pattern Recognition

Assigned Area: Area 12 - Data analysis: clustering, associations mining, classification

ORGANIZERS

• Gabriella Casalino (University of Bari, Italy)
• Giovanna Castellano (University of Bari, Italy)
• Daniel Leite (Federal University of Lavras, Brazil)
• Reinaldo Palhares (Federal University of Minas Gerais, Brazil)

SCOPE

Data stream mining and modeling is a recent methodology that deals with the analysis of potentially large volumes of ordered sequences of data samples. Sensor networks, e-mails, online transactions, network traffic, weather forecasting, health monitoring, industrial process monitoring, and social networks are just some of the most common sources of this kind of data. Stream data arrive continuously. They dynamically change over the time, and need to be processed as soon as they arrive, in a finite amount of time. The idea is to capture the essence of the information within the data, and represent it in the parameters and structure of a model. Thus, special-purpose evolving data analysis methods, which are able to identify patterns in data in quite a real time, are needed to address major challenges such as nonstationarity (concept change) and large datasets (Big data).

TOPICS

This special session is intended to collect novel ideas and share different experiences in the field of soft computing for evolving data streams. Submission of papers covering theoretical and application aspects are encouraged. Topics of interest include, but are not limited to:

• Evolving Soft Computing Techniques
• Evolving Fuzzy Systems
• Evolving Rule-Based Classifiers
• Evolving Neuro-Fuzzy Systems
• Fuzzy techniques for data stream mining
• Learning in non-stationary environments
• Online/incremental Fuzzy clustering
• Adaptive Fuzzy Systems
• Online Genetic and Evolutionary Algorithms
• Adaptive Pattern Recognition
• Incremental and Evolving Fuzzy ML Classifiers
• Big Data analysis through Fuzzy techniques
• Real-time pattern recognition
• Fuzzy techniques for Diagnostics and Prognostics
• Applications: Anomaly Detection, Text-mining, eHealth, Economics, Computer Vision, Process Mining, Internet of Things, Industry 4.0, Learning Analytics, Cyber Security, etc.

SS5 - Data Management in IoT-based Smart Environments:
Leveraging Computational Intelligence for Decision Making (DaMISE’2021)

Assigned Area: Area 16 - Engineering applications incl. IoT, environmental modeling, bioscience engineering

ORGANIZERS

• Allel Hadjali (LIAS/ENSMA, Poitiers, France)
• Miroslav Hudec (University of Economics in Bratislava, Slovakia)
• Edy Portmann (University of Fribourg, Switzerland)

SCOPE

Internet of Things (IoT) is a paradigm that connects multiple and diverse smart objects via Internet. Nowadays, this paradigm is receiving a momentous interest in a number of real-life fields including industry, transport, healthcare and smart cities. This interest will be more and more growing in the future due to the unprecedented number of objects/devices that will be connected in the world. In www.statista.com, this number has been estimated to more than 75 billion devices by 2025. It is expected to rich 125 billion by 2030. Interconnected smart objects will then become the major data producers and consumers instead of humans and they generate tremendous amounts of data using their sensors every single of second. Such IoT data are inherently uncertain, erroneous and noisy on the one hand, and voluminous, distributed and continuous on the other hand.

On the other hand, a smart environment is a connected small world where sensor-enabled connected devices work collaboratively to make the lives of persons comfortable, the business of enterprises much big and flourishing, and so on. It is capable of obtaining knowledge and applying it to satisfy more complex users’ needs.

Recent research efforts have been conducted to integrate IoT with smart environments. This integration allows extending the capabilities of smart objects by enabling the user to monitor the environment from remote sites. IoT-based smart environments have two main and unique characteristics: the prediction capabilities and the decision-making capabilities. Such environments can collect a variety of data from different sources (i.e., objects/devices) and apply data fusion and mining/learning techniques to leverage and analyse data gathered. Hence, data are then the base for making intelligent decisions and providing new services.

Nowadays, in the context of IoT-based smart environments, data management constitutes a modern and a hot topic and has raised many challenging research tasks. New solutions and revisited existing ones are proposed to address such tasks. The main goal of this Special Session is to provide an international forum for researchers from academia and industry to exchange ideas and experiences regarding current and future solutions for managing IoT data. Especially, solutions that leverage techniques borrowed from computational Intelligence field (Soft computing, Fuzzy logic, Uncertainty models, Neural networks, Evolutionary computing, …).

Works in this field should address business, governmental and societal needs. With a holistic objective, it places humans at the epicentre of the digital transformation to unleash our collective potential, thus taking biological as well as artificial intelligence into account. Humans and institutions should have benefits of explained knowledge from the IoT data in the understandable ways, i.e., linguistically. Next, the promising field is focusing on nature inspired approaches in machine learning and computational intelligence for sustainability reasons, as nature-inspired approaches are not only more natural towards humanity but also more energy-efficient than today’s technology, which is the key factor when analysing huge amounts of IoT data. Our emerging, symbiotic technology thus should intend people to support a life in harmony with each other as well as with Nature.

• IoT Data analytics
• IoT Data Fusion / Mining / Integration
• Data quality, Data collection and warehousing
• IoT Data cleaning / visualization
• IoT Data uncertainty management
• Knowledge discovery from IoT Data
• Linguistic summarization and interpretation
• Optimization based on IoT Data
• Energy and time efficient mining knowledge from IoT Data
• Trust, Security and privacy managing
• IoT Data integrity and confidentiality
• Predictive and advanced machine learning models
• IoT-enabled services
• Real-time and stream processing techniques
• Real-life case studies
• Negative aspects and possible solutions

SS6 - Representing and Managing Uncertainty: different scenarios, different tools

Assigned Area - choice a): IJCRS - International Joint Conference on Rough Sets

Assigned Area - choice b): Area 8 - Rough sets and formal concept analysis

Remark: Note that the authors may choose between submitting to IJCRS proceedings (choice a): Springer) or to IFSA-EUSFLAT proceedings (choice b): Atlantis Press).
Follow the instructions when preparing the manuscript.

ORGANIZERS

• Davide Ciucci (University of Milano-Bicocca, Italy)
• Chris Cornelis (Ghent University, Belgium)
• Jesús Medina (University of Cádiz, Spain)
• Dominik Slezak (University of Warsaw, Poland)

SCOPE

The notion of uncertainty has been extensively analysed in the last decades by philosophers, logicians and computer scientists. Here we are interested in the uncertainty originated by different characteristics and flaws in information: incompleteness, imprecision, graduality, granularity, contradiction between agents, etc. For each of these aspects one (or more) specific tool has been introduced in literature: fuzzy sets, rough sets, formal concept analysis, possibility theory, Dempster-Shafer theory, interval analysis, compound objects comparators, etc. Further, when more than one form of uncertainty are present at the same time, it seems natural to fuse such tools, as in the fuzzy rough set case. The special session is devoted to collect all contributions that deal with scenarios leading to a form of uncertainty and tools to represent and manage it. In particular, all critical discussions, comparisons among two or more forms of uncertainty and/or comparisons and fusion of two or more tools are welcome.

The not exhaustive list of topics includes:

• Fuzzy rough sets
• Interval-valued fuzzy sets
• Formal concept analysis
• Mathematical morphology
• Fuzzy relation equations
• Possibility theory
• Dempster-Shafer theory
• Supervaluations
• Near sets
• Interval analysis
• Grey sets
• Non-classical logics (many valued, paraconsistent, epistemic, etc.)
• Similarity-based reasoning
• Networks of comparators

SS7 - Mathematical Fuzzy Logic

Assigned Area: Area 4 - Many-valued logics and algebraic structures for reasoning

ORGANIZERS

• Tommaso Flaminio (IIIA-CSIC, Spain)
• Brunella Gerla (University of Insubria, Italy)
• Lluis Godo (IIIA-CSIC, Spain)
• Amanda Vidal (IIIA-CSIC, Spain)

ABSTRACT

The study of logical, algebraic and proof theoretical tools for the management of imprecision and uncertainty is a well-established line of research, whose development has significantly influenced many areas of applied research, from Economics and Game Theory to Artificial Intelligence. This special session, expression of the EUSFLAT working group of mathematical fuzzy logic, aims at collecting papers about formal approaches to the theories of vagueness, uncertainty and imprecise information management that can be treated in the realm of fuzzy or many-valued logics, and the possible applications in modeling different types of reasoning.

It is focused on (but is not limited to) the following topics:

• Algebraic structures of many-valued logics
• Proof theory of many-valued logics
• Uncertainty theories on many-valued logics
• Modal fuzzy logics and their applications
• Game semantics of many-valued logics
• Topological and categorical methods for many-valued logics

SS8 - Fuzzy Methods in Statistics and Data Analysis

Assigned Area: Area 7 - Uncertainty modeling, fuzzy probability and statistics

ORGANIZERS

• Przemyslaw Grzegorzewski (Warsaw University of Technology and Polish Academy of Sciences, Poland)
• Jin Hee Yoon (Sejong University, South Korea)

SCOPE

The rapidly growing stream of diverse data, increasing complexity of systems explored in science and engineering and new challenges data analysts face reveal the need for more flexible uncertainty modeling tools. Such challenges and needs require soft modeling and computing that are less rigid than traditional approaches and techniques, and therefore easier to adapt to the real nature of the information. The desired methodology should combine different possible types of uncertainty, including randomness, imprecision, ambiguity, etc. By integrating fuzzy sets and probability theory, 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 fuzzy 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
• Clustering and classification
• Fuzzy random variables
• Fuzzy regression methods
• Imprecise probabilities
• Interpretable models
• Interval data
• Machine learning
• Possibility theory
• Random sets
• Robust statistics
• Soft computing
• Statistical software for imprecise data

SS9 - Fuzzy Technologies for Web Intelligence and Internet of Things

Assigned Area: Area 16 - Engineering applications incl. IoT, environmental modeling, bioscience engineering

ORGANIZERS

• Giovanni Acampora (University of Naples Federico II, Italy)
• Chang-Shing Lee (National University of Tainan, Taiwan)
• Amir Pourabdollah (Nottingham Trent University, UK)
• Autilia Vitiello (University of Naples Federico II, Italy)
• Marek Reformat (University of Alberta, Canada)

SCOPE

The constant growth of the Internet and introduction of such concepts as the Semantic Web and Internet of Things create challenges as well as opportunities to transform the internet into an environment providing the users and any internet enabled devices with the abilities to utilize and explore efficient way. Fuzzy Logic and Soft Computing offer techniques and methods suitable for dealing with imprecision, fusing information from multiple sources, selecting best among multiple alternatives, or representing information and knowledge using.

The special session will focus on the current research trends in the area of theory and practical aspects of intelligent systems equipped with fuzzy and other soft computing methods suitable for addressing issues specific to the internet of things and representation and processing of information and knowledge existing on the web.

The topics of particular interest to the session include but are not limited to:

• multi-criteria decision-making
• internet of things
• information fusion
• semantic-based processing of data
• approximate reasoning
• fuzzy ontology and ontology-based systems
• knowledge- and rule-based systems
• hybrid intelligent systems
• recommendation systems
• context-aware systems
• information retrieval and knowledge discovery

SS10 - Modelling Vagueness in Natural Language

Assigned Area: Area 5 - Generalized quantifiers and natural language modeling

ORGANIZERS

• Vilém Novák (University of Ostrava, Czech Republic)
• Adrià Torrens Urrutia (University of Ostrava, Czech Republic)
• María Dolores Jiménez López (Universitat Rovira i Virgili, Spain)

SCOPE

This special session is equally addressed to linguists, computational linguists, mathematicians and computer scientists, who have an interest in modelling natural language. Therefore, the main objective of this special session is bringing together researchers who study language from different areas. We want to encourage discussion from different perspectives when it comes to model the inherent vagueness of natural languages.

Vagueness is a conception concerning with those objects which are difficult to be classified categorically at first sight. Therefore, "modelling vagueness" is equivalent to studying linguistic objects with a non-discrete approach. There are many objects in natural languages which are prone to poly-signification, and they are essentially context-dependent.

TOPICS

In this special session, we encourage proposals from different fields of science which tackle the necessity to deal with the vague phenomena in language:

• Artificial Intelligence
• Computational Social Science and Social Media
• Dialog and Interactive Systems
• Discourse and Pragmatics
• Formal linguistics
• Information Extraction and Text Mining
• Information Retrieval, Search and Question Answering
• Interpretable models
• Interpretability and Analysis of Models for NLP
• Language Resources and Evaluation
• Language Grounding to Vision, Robotics and Beyond
• Linguistic Theories, Cognitive Modeling and Psycholinguistics
• Machine Learning, Deep learning, and Neural Networks for NLP
• Machine Translation
• Multilinguality
• Possibility theory
• Linguistic Semantics
• Sentiment Analysis, Stylistic Analysis, and Argument Mining
• Speech
• Syntax

SS11 - Models and Proofs of Cluster Validation

Assigned Area: Area 12 - Models and Proofs of Cluster Validation

ORGANIZERS

• A Mani (Indian Statistical Institute, Kolkata, India)
• Sushmita Mitra (Indian Statistical Institute, Kolkata, India)

ABSTRACT

It is well-known that concepts of cluster validity and associated rationale are far from being satisfactory from multiple perspectives. Indices are usually controversial. The scope of a model based approach from a statistical, mereo-topological, graph-theoretical, neural network or soft perspective, for example, is apparently limited by methodological constraints and implicit ontology. The purpose of this special session is to improve the dialogue between researchers from diverse fields specifically by way of (though not limited to) research papers and surveys on interconnections, novel methodologies, models, connections with logic, the boundaries of semi-supervised techniques, and the very idea of proof in a clustering context.

TOPICS

Topics of interest include all the following (but are not limited to):

• Models of Hard and Soft Clustering (Mathematical, Statistical, Logical, Mereological, and Computational)
• Ontology in Applications (including Big Data and Sociological contexts)
• Comparative Studies on Indices
• Proofs and Truth in Clustering
• Spectral, Graph Theoretic, and Hierarchical Clustering
• Rough, Fuzzy, Hybrid, and Soft Clustering
• Feature Selection and Extraction for Clustering, and Related Ontology
• Re-sampling Methods
• Image Analysis, and Visual Clustering

SS12 - Fuzzy Modelling and Control of Dynamic Systems

Assigned Area: Area 14 - Fuzzy Control

ORGANIZERS

• Juan Manuel Escaño (Universidad de Sevilla, Spain)
• Federico Cuesta (Universidad de Sevilla, Spain)
• Juan Albino Méndez (University of La Laguna, Spain)

DESCRIPTION

Intelligent control systems cover different techniques such as expert systems, fuzzy logic, neural networks, genetic algorithms, etc. However, fuzzy logic has proven to be the most widely used technique in industrial applications. The keys to this are its ability to represent expert control/modeling knowledge and experience and its ease of implementation. Another factor contributing to the deployment of fuzzy applications in industry is the increasing computational power of embedded systems. Thus, fuzzy inference systems are present at all levels of the automation pyramid and appear as a powerful tool for modeling, identifying and controlling complex plants.

The scope of the session will focus on the modeling and/or control of dynamic systems, using fuzzy inference systems, including synergies with other techniques. Both relevant theoretical contributions and applications will be considered. The objective of this initiative is to bring together researchers from the field of intelligent systems who focus their interest on modeling and/or control using fuzzy logic, and gather their latest advances in their state of the art or applications. As such, the session aims to provide an overview of current high-quality research in this area, while at the same time exploring new trends in fuzzy modeling and control methodologies.

TOPICS

Topics include, but are not limited to:

• Fuzzy control systems
• Fuzzy identification and fuzzy modeling
• Hybrid-fuzzy modelling and identification
• Fuzzy clustering models
• Fuzzy modelling of high dimensional systems
• Inverse model identification

SS13 - Explainable AI

Assigned Area: Area 10 - Machine learning and AI

ORGANIZERS

• Kelly Cohen (University of Cincinnati, USA)
• Vladik Kreinovich (University of Texas at El Paso, USA)

ABSTRACT

Many systems use machine learning to make recommendations -- e.g., whether to give a loan. However, machine learning systems are imperfect, and since they do not explain anything, it's difficult to detect erroneous decisions. Many countries – including EU – require that AI-based decisions be accompanied by natural-language explanations. To achieve this goal, it is reasonable to use known techniques relating natural-language knowledge and precise decisions, such as fuzzy logic. The purpose of this session is to describes results and ideas about using fuzzy techniques to make AI-based decisions explainable.

The goal of the special session is to focus mainly (not exclusively) on the following three crucial questions and try to answer them at least partially by the accepted talks:

• How can we use fuzzy techniques to make existing controls and recommendations (e.g., recommendations obtained from deep learning) explainable?
• How can we train fuzzy systems to provide accurate and explainable alternatives to traditional machine learning techniques?
• How can we use fuzzy logic to incorporate fairness and equity into decision-making systems?

SS14 - CI-4-SDG: Computational Intelligence to help reaching the Sustainable Development Goals

Assigned Area: Area 16 - Engineering applications incl. IoT, environmental modeling, bioscience engineering

ORGANIZERS

• David A. Pelta (Universidad de Granada, Spain)
• Enrique Onieva (Universidad de Deusto, Spain)
• Antonio Masegosa (Universidad de Deusto, Spain)

ABSTRACT

The Sustainable Development Goals are a universal call to action to end poverty, protect the planet and improve the lives and prospects of everyone, everywhere. The 17 Goals were adopted by all UN Member States in 2015, as part of the 2030 Agenda for Sustainable Development which set out a 15-year plan to achieve the Goals (see https://sdgs.un.org/ ).

At the core of such goals there are a hughe number of topics related with: intelligent transport, services deployment, health promotion, smart societies, disaster and crisis management, etc. In each of these challenges it is possible to recognize decision and optimization problems leading to models and frameworks (be it mathematical, linguistic, computational…) requiring suitable models and solution methods.

Uncertainty, vagueness, imprecision, dynamism, etc. are ubiquitous in this context. These features should not be ignored and should be properly managed. It is here where Fuzzy Sets and Systems, and hence Computational Intelligence (CI) based methodologies and techniques, can help to model and solve (at least partly) some problems related with the SDGs.

The aim of this special session is to serve as a meeting point and discussion forum for researchers and practitioners on the latest developments on CI to help reaching the SDGs. In this context, we will welcome both theoretical and more application oriented contributions addressing

SDGs related problems in the following (but not limited) areas:

• Intelligent Transport Systems
• Renewable Energies
• Logistics
• Health
• Agriculture

SS15 - Modeling Decisions in AI

Assigned Area: Area 9 - Fuzzy systems: decision-making, inference systems, preference modeling, optimization

ORGANIZERS

• Camilo Franco (University of the Andes, Colombia)
• J. Tinguaro Rodriguez (University Complutense of Madrid, Spain)
• Javier Montero (University Complutense of Madrid, Spain)
• Daniel Gómez (University Complutense of Madrid, Spain)
• Jens Leth Hougaard (University of Copnhagen, Denmark)

ABSTRACT

The purpose of the special session is to gather researchers in the area of decision modeling in AI. Decision modeling is a fundamental field for explaining and solving real world problems, where imprecision and vagueness emerge as key aspects for properly representing the knowledge contained in common sources of information. Besides representing human reasoning processes for building automatic decision support, AI algorithms should consider their own explainability and fairness of their final output.

Original research works dealing with but not limited to the following topics are welcome: fuzzy preference modelling, knowledge and uncertainty representation, multicriteria decision making, interpretability/explainability of machine learning algorithms, algorithmic decision making and algorithmic fairness.

SS16 - Mathematical Methods Towards Dealing with Uncertainty in Data Sciences

Assigned Area: Area 6 - Dynamical modeling: fuzzy dynamical systems and fuzzy differential equations

ORGANIZERS

• Irina Perfilieva (University of Ostrava, Czech Republic)
• Michal Holčapek (University of Ostrava, Czech Republic)
• Anand Pratap Singh (University of Ostrava, Czech Republic)

ABSTRACT

It is well known that the use of techniques from fuzzy set theory is very effective when we deal with the problems having uncertain parameters. The aim of this special session is to discuss mathematical methods that are focused on elaboration various structured spaces characterizing uncertainty in many of its facets. Another focus is to show that a certain amount of uncertainty is useful in finding weak (robust) solutions to many classical problems in applied fields, connected with dynamic processes, that are modelled by differential, integral, stochastic equations, and their fuzzy versions.

We solicit contributions that show how sophisticated theories contribute to non-trivial solutions to problems in data sciences including those that are ill-defined or have non-standard solutions.

Topics of this special session include but are not limited to:

• Fuzzy integral and differential equations
• Fuzzy automata
• Fuzzy dynamical systems
• Fuzzy transforms and applications
• Uncertainity in relational structures
• Fuzzy topology

SS17 - Theoretical Aspects of Pre-Aggregation Functions and Generalized Forms of Monotonicity and Aggregation

Assigned Area: AGOP - Summer School on Aggregation Operators

ORGANIZERS

• Humberto Bustince (Universidad Pública de Navarra, Spain)
• Javier Fernández (Universidad Pública de Navarra, Spain)
• Graçaliz Dimuro (Universidade Federal do Rio Grande, Brazil & Universidad Pública de Navarra, Spain)

ABSTRACT

This special session is aimed at discussing the most recent theoretical developments related to fusion functions satisfying some generalized forms of monotonicity. In particular, the session in dedicated to weaker forms of monotonicity, e.g., directional monotonicity and ordered directional monotonicity. Then, the session will focus on generalizations of the concept of aggregation functions when one consider some generalized forms of monotonicity and/or boundary conditions. Pre-aggregation functions have appeared in the literature in 2016 and, since then, the interest on new classes of functions derived by pre-aggregation functions has increased a lot. Such functions encompass both classical aggregation functions and other weaker functions that do not fulfil the full monotonicity condition, but present excellent behavior in aggregation processes, so offering more flexibility in applications. In this sense, the objective of this session is to provide researchers in the field with an opportunity to present their most recent developments and to discuss recent trends in this area, as well as to identify potential problems of interest for researchers. The following is a non-exhaustive list of topics that this session intends to cover: Directional monotonicity, Weak monotonicity, Ordered directional monotonicity, Pre-aggregation functions, generalizations of aggregation functions, aggregation in the interval-valued context, generalizations of the Choquet integral, d-Choquet integral.

SS18 - Applications of Pre-Aggregation Functions and Generalized Forms of Aggregation

Assigned Area: AGOP - Summer School on Aggregation Operators

ORGANIZERS

• Humberto Bustince (Universidad Pública de Navarra, Spain)
• Graçaliz Dimuro (Universidade Federal do Rio Grande, Brazil & Universidad Pública de Navarra, Spain)
• Javier Fernández (Universidad Pública de Navarra, Spain)

ABSTRACT

This special session is dedicated to discuss the potential applications of pre-aggregation functions and other classes of aggregation-like functions satisfying weaker monotonicity/boundary conditions, including the interval-valued context, e.g., directional monotonicity and ordered directional monotonicity. The main concept in the field, namely, pre-aggregation functions, have appeared in the literature in 2016, in the context of classification problems, replacing classical aggregation operators used in the fuzzy reasoning mechanism of fuzzy rule-based systems (FRBS). The excellent performance provided by pre-aggregation functions in FRBSs have increased the interest of the researchers in their use in other kinds of applications that require some kind aggregation process but the full standard monotonicity may be not required, like image processing and deep learning. Also, their use in real classification problems are appearing in the literature. Then, the objective of this session is to provide researchers in the field with an opportunity to present their most recent developments in applications and to discuss recent trends in this area, as well as to identify potential application problems of interest for researchers. The following is a non-exhaustive list of topics that this session intends to cover: Applications of pre-aggregation functions and generalized forms of aggregation in classification problems, image processing, deep learning, decision making, real case studies.

SS19 - Deep Learning Under Uncertain and Imperfect Information

Assigned Area: Area 10 - Machnie learning and AI

ORGANIZERS

• Humberto Bustince (Universidad Pública de Navarra, Spain)
• Javier Fernández (Universidad Pública de Navarra, Spain)
• Graçaliz Dimuro (Universidade Federal do Rio Grande, Brazil & Universidad Pública de Navarra, Spain)

ABSTRACT

Deep learning is concerned with the field within the world of machine learning that is inspired by the working mechanisms of the human brain. This objective is pursued through the concept of artificial neural networks, computational models that seek to learn hierarchical representations of the input data, through the sequential composition of non-linear functions. In recent years, applications of these models have seen an unprecedented expansion, assuming the state of the art in fields such as image processing, reinforced learning or natural language processing, among others. This special session is aimed at discussing the most recent developments related to Deep Learning considering uncertain and imperfect information. In particular, this special session will discuss the role of aggregation functions in the process of deep learning, fuzzy neural networks, reduction of dimensionality with fuzzy techniques, the use of intervals in deep learning, manifold Learning, transfer learning, embeddings and projections, as well applications in general.