Advances in Hybridization of Intelligent Methods: Models, Systems and Applications PDF
Preview Advances in Hybridization of Intelligent Methods: Models, Systems and Applications
Smart Innovation, Systems and Technologies 85 Ioannis Hatzilygeroudis Vasile Palade Editors Advances in Hybridization of Intelligent Methods Models, Systems and Applications 123 Smart Innovation, Systems and Technologies Volume 85 Series editors Robert James Howlett, Bournemouth University and KES International, Shoreham-by-sea, UK e-mail: [email protected] Lakhmi C. Jain, University of Canberra, Canberra, Australia; Bournemouth University, UK; KES International, UK e-mails: [email protected]; [email protected] About this Series The Smart Innovation, Systems and Technologies book series encompasses the topics of knowledge, intelligence, innovation and sustainability. The aim of the seriesistomakeavailableaplatformforthepublicationofbooksonallaspectsof single and multi-disciplinary research on these themes in order to make the latest resultsavailableinareadily-accessibleform.Volumesoninterdisciplinaryresearch combining two or more of these areas is particularly sought. The series covers systems and paradigms that employ knowledge and intelligence in a broad sense. Its scope is systems having embedded knowledge and intelligence, which may be applied to the solution of world problems in industry, the environment and the community. It also focusses on the knowledge-transfer methodologies and innovation strategies employed to make this happen effectively. The combination of intelligent systems tools and a broad range of applications introduces a need for a synergy of disciplines from science, technology, business and the humanities. The series will include conference proceedings, edited collections, monographs, handbooks, reference books, and other relevant types of book in areas of science and technology where smart systems and technologies can offer innovative solutions. Highqualitycontentisanessentialfeatureforallbookproposalsacceptedforthe series. It is expected that editors of all accepted volumes will ensure that contributionsaresubjectedtoanappropriatelevelofreviewingprocessandadhere to KES quality principles. More information about this series at http://www.springer.com/series/8767 ⋅ Ioannis Hatzilygeroudis Vasile Palade Editors Advances in Hybridization of Intelligent Methods Models, Systems and Applications 123 Editors Ioannis Hatzilygeroudis Vasile Palade Schoolof Engineering, Department TheFaculty ofEngineering, Environment ofComputerEngineeringandInformatics andComputing,Schoolof Computing University of Patras CoventryUniversity Patras Coventry Greece UK ISSN 2190-3018 ISSN 2190-3026 (electronic) Smart Innovation,Systems andTechnologies ISBN978-3-319-66789-8 ISBN978-3-319-66790-4 (eBook) https://doi.org/10.1007/978-3-319-66790-4 LibraryofCongressControlNumber:2017950251 ©SpringerInternationalPublishingAG2018 Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpart of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission orinformationstorageandretrieval,electronicadaptation,computersoftware,orbysimilarordissimilar methodologynowknownorhereafterdeveloped. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publicationdoesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfrom therelevantprotectivelawsandregulationsandthereforefreeforgeneraluse. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authorsortheeditorsgiveawarranty,expressorimplied,withrespecttothematerialcontainedhereinor for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictionalclaimsinpublishedmapsandinstitutionalaffiliations. Printedonacid-freepaper ThisSpringerimprintispublishedbySpringerNature TheregisteredcompanyisSpringerInternationalPublishingAG Theregisteredcompanyaddressis:Gewerbestrasse11,6330Cham,Switzerland Preface The invention of hybrid intelligent methods is a very active research area in arti- ficial intelligence (AI). The aim is to create hybrid methods that benefit from each of their components. It is generally believed that complex problems can be easily solved with hybrid methods. By “hybrid,” we mean any kind of combined use (either tight or loose) of distinct intelligent methods toward solving a problem, eitherspecificorgeneral.Inthissense,itisusedassynonymouswith“integrated.” Some of the existing efforts try to make hybrids of what are called soft com- putingmethods(fuzzylogic,neuralnetworks,andgeneticalgorithms)eitheramong themselves or with more traditional AI methods, such as logic and rules. Another stream of efforts integrates case-based reasoning or machine learning with soft computing or traditional AI methods. Yet another integrates agent-based approa- cheswithlogicandnon-symbolicapproaches.Someofthecombinationshavebeen quite important and more extensively used, like neuro-symbolic methods, neuro-fuzzy methods, and methods combining rule-based and case-based reason- ing.However,thereareothercombinationsthatarestillunderinvestigation,suchas those related to the Semantic Web and Big Data areas. For example, the recently emergeddeeplearningarchitecturesormethodsarealsohybridbynature.Insome cases,integrationsarebasedonfirstprinciples,creatinghybridmodels,whereasin othercasestheyarecreatedinthecontextofsolvingproblemsleadingtosystemsor applications. Important topics of the above area are (but not limited to) the following: (cid:129) Case-Based Reasoning Integrations (cid:129) Ensemble Learning, Ensemble Methods (cid:129) Evolutionary Algorithms Integrations (cid:129) Evolutionary Neural Systems (cid:129) Fuzzy-Evolutionary Systems (cid:129) Semantic Web Technologies Integrations (cid:129) Hybrid Approaches for the Web (cid:129) Hybrid Knowledge Representation Approaches/Systems (cid:129) Hybrid and Distributed Ontologies v vi Preface (cid:129) Information Fusion Techniques for Hybrid Intelligent Systems (cid:129) Integrations of Neural Networks (cid:129) Integrations of Statistical and Symbolic AI Approaches (cid:129) Intelligent Agents Integrations (cid:129) Machine Learning Combinations (cid:129) Neuro-Fuzzy Approaches/Systems (cid:129) Swarm Intelligence Methods Integrations (cid:129) Applications of Combinations of Intelligent Methods to – Biology and Bioinformatics – Education and Distance Learning – Medicine and Health Care – Multimodal Human–Computer Interaction – Natural Language Processing and Understanding – Planning, Scheduling, Search, and Optimization – Robotics – Social Networks This volume includes extended and revised versions of some of the papers presented in the 6th International Workshop on Combinations of Intelligent Methods and Applications (CIMA 2016) and also papers submitted especially for this volume after a CFP. CIMA 2016 was held in conjunction with the 22nd EuropeanConferenceonArtificialIntelligence(ECAI2016),August30,2016,The Hague, Holland. Papers went through a peer review process by the CIMA-16 program committee members. Giannopoulos et al. present results on using two deep learning methods (Goo- gleNet and AlexNet) on facial expression recognition. The paper of Haque et al. presentsresultsonhowcommunicationmodelaffectsroboticsswarmperformance. Jabreel et al. introduce and experiment with a target-dependent sentiment analysis approachfortweets.Maniaketal.presentahybridapproachusedforthemodeling andpredictionoftaxiusageinthecontextofsmartcities.ThepaperofMasonetal. introducesareinforcementlearningapproachcombiningaMarkovdecisionprocess andquantificationverificationtorestrict an agent’sbehavioratasafe level.Mosca and Magoulas propose a method for approximating an ensemble of deep neural networks by asingledeep neural network. Finally,Teppan andFriedrichpresenta constraint answer programming solver and investigate its performance through its application to two manufacturing problems. Wewouldliketoexpressourappreciationtoalltheauthorsofsubmittedpapers as well as to the members of CIMA 2016 program committee for their excellent review work. Wehopethatthiskindofpost-proceedingswillbeusefultobothresearchersand developers. Patras, Greece Ioannis Hatzilygeroudis Coventry, UK Vasile Palade Reviewers (From CIMA 2016 Program Committee) Plamen Agelov, Lancaster University, UK Nick Bassiliades, Aristotle University of Thessaloniki, Greece Kit Yan Chan, Curtin University, Australia Gloria Cerasela Crisan, Vasile Alecsandri University of Bacau, Romania Georgios Dounias, University of the Aegean, Greece Foteini Grivokostopoulou, University of Patras, Greece Ioannis Hatzilygeroudis, University of Patras, Greece (Co-chair) Andreas Holzinger, TU Graz and MedUni Graz, Austria George Magoulas, Birkbeck College, UK Christos Makris, University of Patras, Greece Antonio Moreno, University Rovira i Virgili, Spain Vasile Palade, Coventry University, UK (Co-chair) Isidoros Perikos, University of Patras, Greece Camelia Pintea, Technical University of Cluj-Napoca, Romania Jim Prentzas, Democritus University of Thrace, Greece Roozbeh Razavi-Far, Politecnico di Milano, Italy David Sanchez, University Rovira i Virgili, Spain Kyriakos Sgarbas, University Of Patras, Greece Douglas Vieira, Enacom-Handcrafted Technologies, Brazil vii Contents Deep Learning Approaches for Facial Emotion Recognition: A Case Study on FER-2013 .. ..... .... .... .... .... .... ..... .... 1 Panagiotis Giannopoulos, Isidoros Perikos and Ioannis Hatzilygeroudis Analysis of Biologically Inspired Swarm Communication Models.. .... 17 Musad Haque, Electa Baker, Christopher Ren, Douglas Kirkpatrick and Julie A. Adams Target-Dependent Sentiment Analysis of Tweets Using Bidirectional Gated Recurrent Neural Networks . .... .... .... .... .... ..... .... 39 Mohammed Jabreel, Fadi Hassan and Antonio Moreno Traffic Modelling, Visualisation and Prediction for Urban Mobility Management .. .... .... .... ..... .... .... .... .... .... ..... .... 57 Tomasz Maniak, Rahat Iqbal and Faiyaz Doctor Assurance in Reinforcement Learning Using Quantitative Verification ... .... .... .... ..... .... .... .... .... .... ..... .... 71 George Mason, Radu Calinescu, Daniel Kudenko and Alec Banks Distillation of Deep Learning Ensembles as a Regularisation Method... .... .... .... .... ..... .... .... .... .... .... ..... .... 97 Alan Mosca and George D. Magoulas Heuristic Constraint Answer Set Programming for Manufacturing Problems . ..... .... .... .... .... .... ..... .... 119 Erich C. Teppan and Gerhard Friedrich ix Deep Learning Approaches for Facial Emotion Recognition: A Case Study on FER-2013 Panagiotis Giannopoulos, Isidoros Perikos and Ioannis Hatzilygeroudis Abstract Emotions constitute an innate and important aspect of human behavior that colors the way of human communication. The accurate analysis and interpre- tation of the emotional content of human facial expressions is essential for the deeper understanding of human behavior. Although a human can detect and interpretfacesandfacialexpressionsnaturally,withlittleornoeffort,accurateand robust facial expression recognition by computer systems is still a great challenge. The analysis of human face characteristics and the recognition of its emotional statesareconsideredtobeverychallenginganddifficulttasks.Themaindifficulties comefromthenon-uniformnatureofhumanfaceandvariationsinconditionssuch as lighting, shadows, facial pose and orientation. Deep learning approaches have been examined as a stream of methods to achieve robustness and provide the necessary scalability on new type of data. In this work, we examine the perfor- manceoftwoknowndeeplearningapproaches(GoogLeNetandAlexNet)onfacial expression recognition, more specifically the recognition of the existence of emo- tional content, and on the recognition of the exact emotional content of facial expressions. The results collected from the study are quite interesting. ⋅ ⋅ Keywords Affective computing Facial emotion recognition Deep ⋅ ⋅ ⋅ learning Convolutional neural networks GoogLeNet Alexnet P.Giannopoulos(✉) ⋅ I.Perikos ⋅ I.Hatzilygeroudis DepartmentofComputerEngineeringandInformatics, UniversityofPatras,Patras,Greece e-mail:[email protected] I.Perikos e-mail:[email protected] I.Hatzilygeroudis e-mail:[email protected] I.Perikos TechnologicalEducationalInstituteofWesternGreece,Patras,Greece ©SpringerInternationalPublishingAG2018 1 I.HatzilygeroudisandV.Palade(eds.),AdvancesinHybridization ofIntelligentMethods,SmartInnovation,SystemsandTechnologies85, https://doi.org/10.1007/978-3-319-66790-4_1
The list of books you might like
