IEEE Computational Intelligence Society, France Chapter
Home > About
Due to the covid19 virus, we have to postpone the workshop until further notice. Our apologies.
WORKSHOP ON COMPUTATIONAL INTELLIGENCE
the IEEE Computational Intelligence Society,
the IEEE France Section CIS Chapter
The underside of the Eiffel Tower at night
Location: LIP6, Sorbonne Université
4 place Jussieu, 75005 Paris
Tower 26, floor 1, corridor 25-26, Room 105
Free attendance, registration at https://app.smartsheet.com/b/form/8e8e02f9244a4f03870d2564ccc5a0fc
Introduction to the Workshop on Computational Intelligence
IEEE CIS President
LIP6, Paris, France
bernadette.bouchon-meunier [ at ] lip6.fr
Bernadette Bouchon-Meunier is a director of research emeritus at the National Centre for Scientific Research in the Computer Science Laboratory (LIP6) of Sorbonne Université. She was the head of the department of Databases and Machine Learning in the Computer Science Laboratory of the University Pierre et Marie Curie-Paris 6. She is the Editor-in-Chief of the International Journal of Uncertainty, Fuzziness and Knowledge-based Systems, the (co)-editor of 27 books, and the (co)-author of five. She has (co)-authored more than 400 papers on approximate and similarity-based reasoning, as well as the application of fuzzy logic and machine learning techniques to decision-making, data mining, risk forecasting, information retrieval, user modelling, sensorial and emotional information processing. She was the Vice-President for Conferences of the IEEE Computational Intelligence Society (2014-2018), the President-Elect of the CIS in 2019 and the Vice-President for Chapters of the IEEE France Section (2014-2019). She is currently the President of the IEEE Computational Intelligence Society and the IEEE France Section Computational Intelligence Chapter Vice-Chair. She is an IEEE Life Fellow, an International Fuzzy Systems Association Fellow and an Honorary Member of the EUSFLAT Society. She received the 2012 IEEE Computational Intelligence Society Meritorious Service Award, the 2017 EUSFLAT Scientific Excellence Award and the 2018 IEEE Computational Intelligence Society Fuzzy Systems Pioneer Award.
AI (via CI often “ignoring” BI): What We Might Be Missing – a personal account
Nikhil R. Pal
IEEE CIS Past President
Center for Artificial intelligence and Machine Learning
Electronics and communication Sciences Unit
Indian statistical Institute, Calcutta, India
nrpal59 [ at ] gmail.com
We are in the era of Artificial Intelligence (AI), in particular, Artificial Narrow Intelligence. AI has sailed through several “ ups and downs” and has come to the present state. At present we have been witnessing numerous success stories of AI, often beating human performance, and this has caused our expectation from AI to skyrocket. In many cases, neural networks (a major component of Computational Intelligence (CI)), in particular deep neural networks, recurrent neural networks, are the main pillars of such systems. Often AI systems ignore biological intelligence (BI). It seems, implicitly we have started believing in philosophies like “bigger the better” (bigger data sets or massive architecture with millions of free parameters) and “data say all”. Such approaches have been proved to be useful but raise some concerns too! In this talk I shall discuss, what in my personal view, are some important issues that need attention (Caveat: one can happily disagree and dissent). And if time permits, I shall briefly describe some of our attempts to “marginally address” some of these issues.
Nikhil R. Pal is a Professor in the Electronics and Communication Sciences Unit of the Indian Statistical Institute. His current research interest includes brain science, computational intelligence, machine learning and data mining. He was the Editor-in-Chief of the IEEE Transactions on Fuzzy Systems for the period January 2005-December 2010. He has served/been serving on the editorial /advisory board/ steering committee of several journals including the International Journal of Approximate Reasoning, Applied Soft Computing, International Journal of Neural Systems, Fuzzy Sets and Systems, IEEE Transactions on Fuzzy Systems and the IEEE Transactions on Cybernetics. He is a recipient of the 2015 IEEE Computational Intelligence Society (CIS) Fuzzy Systems Pioneer Award, He has given many plenary/keynote speeches in different premier international conferences in the area of computational intelligence. He has served as the General Chair, Program Chair, and co-Program chair of several conferences. He was a Distinguished Lecturer of the IEEE CIS (2010-2012, 2016-2018.) and was a member of the Administrative Committee of the IEEE CIS (2010-2012). He has served as the Vice-President for Publications of the IEEE CIS (2013-2016), President of the IEEE CIS (2018-2019) and he is currently Past President of the CIS. He is a Fellow of the National Academy of Sciences, India, Indian National Academy of Engineering, Indian National Science Academy, International Fuzzy Systems Association (IFSA), The World Academy of Sciences, and a Fellow of the IEEE, USA.
Understanding the decisions of a machine: interpretable and explainable AI
IEEE CIS Vice President for Technical Activities
Universidad Politecnica de Madrid, Spain
luis.magdalena [ at ] upm.es
Nowadays, our daily life is strongly conditioned by algorithmic decisions; decisions taken by machines that work on the basis of different AI techniques. Those decisions can even affect our fundamental rights, and consequently, it makes sense to analyse why they were taken. In fact, some recent regulations are considering this as a right, the "right to an explanation" when we are affected by one of these decisions. But in most cases, "deciphering" the reasons supporting an algorithmic decision is not a straightforward task. This is because many of the considered AI algorithms are black-box models.
The talk will consider this question and analyse different approaches that are applied when we need to understand the decision of a machine. These approaches go from the use of intrinsically interpretable models (as fuzzy rules or decision trees), to the superposition of explanation abilities over a model that was designed as a black-box.
Luis Magdalena is with the Dept. of Applied Mathematics for ICT of the Universidad Politecnica de Madrid. From 2006 to 2016 he was Director General of the European Centre for Soft Computing in Asturias (Spain). Under his direction, the Center was recognized with the IEEE-CIS Outstanding Organization Award in 2012.
Prof. Magdalena has been actively involved in more than fifty research projects. He has co-authored or co-edited ten books including “Genetic Fuzzy Systems,” “Accuracy Improvements in Liguistic Fuzzy Modelling,” and “Interpretability Issues in Fuzzy Modeling”. He has also authored over one hundred and seventy papers in books, journals and conferences. His research interests are related to Computational Intelligence theory and applications, including Fuzzy Logic, Genetic Algorithms, Hybrid Intelligent Systems, and their industrial applications.
Prof. Magdalena has been President of the “European Society for Fuzzy Logic and Technologies”, Vice-president of the International Fuzzy Systems Association, and is Vice-President for Technical Activities of the IEEE Computational Intelligence Society for the period 2020-21.
Making sense out of activity sensing in eldercare
IEEE CIS Vice President for Publications
Electrical Engineering and Computer Science Department
University of Missouri-Columbia, USA
KellerJ [ at ] missouri.edu
With the increase in the population of older adults around the world, a significant amount of work has been done on in-home sensor technology to aid the elderly age independently. However, due to the large amounts of data generated by the sensors, it takes a lot of effort and time for the clinicians to makes sense of this data. In this talk, I will survey two connected approaches to provide explanations of these complex sensor patterns as they relate to senior health. Abnormal sensor patterns produced by certain resident behaviors could be linked to early signs of illness. In seven eldercare facilities around Columbia, MO operated by Americare, we have deployed an intelligent elderly monitoring system with summarization and symptom suggesting capabilities for 3 years.
The first procedure starts by identifying important attributes in the sensor data that are relevant to the health of the elderly. We then develop algorithms to extract these important health related features from the sensor parameters and summarize them in natural language, with methods grounded in fuzzy set theory. We focus on making the natural language summaries to be informative, accurate and concise, and have conducted numerous surveys of experts to validate our choices.
The second approach is a framework for detecting health patterns utilizing sensor sequence similarity and natural language processing (NLP). A context preserving representation of daily activities is used to measure the similarity between the sensor sequences of different days. The medical concepts are extracted from the nursing notes, and allow for imputation of potential reasons for health alerts based on the similarity. Joining these two approaches provide a powerful XAI description of early illness recognition for elders.
James M. Keller holds the University of Missouri Curators Distinguished Professorship in the Electrical Engineering and Computer Science Department on the Columbia campus. His research interests center on computational intelligence with a focus on problems in computer vision, pattern recognition, and information fusion including bioinformatics, spatial reasoning, geospatial intelligence, landmine detection and technology for eldercare. Professor Keller has been funded by a variety of government and industry organizations and has coauthored over 500 technical publications.
James M. Keller is a Life Fellow of the IEEE, is an IFSA Fellow, and a past President of NAFIPS. He received the 2007 Fuzzy Systems Pioneer Award and the 2010 Meritorious Service Award from the IEEE Computational Intelligence Society. He finished a full six year term as Editor-in-Chief of the IEEE Transactions on Fuzzy Systems, followed by being the Vice President for Publications of the IEEE CIS from 2005-2008, and then an elected CIS Adcom member. He is VP Pubs for CIS again, and has served as the IEEE TAB Transactions Chair and as a member of the IEEE Publication Review and Advisory Committee from 2010 to 2017. James M. Keller has had many conference positions and duties over the years.
Big Data Challenges and Deep Learning Applications to Astronomy
Pablo A. Estevez
IEEE CIS Vice President for Finances
Department of Electrical Engineering
University of Chile and Millennium Institute of Astrophysics, Chile
paestevez [ at ] gmail.com
Astronomy is facing a paradigm shift caused by the exponential growth of the sample size, data complexity and data generation rates of new sky surveys. To cope with a change of paradigm to data-driven science new computational intelligence approaches are needed. In this talk I will introduce the general context and the big data challenges that astronomy is facing. I will present the ALERCE broker, our state of the art project. Our pipeline includes an early classifier, a late classifier and an outlier detector. I will describe these three components and in particular some deep learning applications. The first application is the early detection of supernovae in astronomical images. We developed a convolutional neural network to discriminate between four types of transients and bogus events. This is implemented in an online platform called SNe Hunter, which is open to the general public. The second application is to classify different classes of astronomical objects based on a recurrent convolutional neural network, which uses directly sequences of images in an online fashion. The late classifier is using hierarchical random forests, and the outlier detector is based on geometric transformations.
Pablo A. Estévez received his professional title in electrical engineering (EE) from Universidad de Chile, in 1981, and the M.Sc. and Dr.Eng. degrees from the University of Tokyo, Japan, in 1992 and 1995, respectively. He is a Full Professor with the Electrical Engineering Department, University of Chile, and former Chairman of the Electrical Engineering Department in the period 2006-2010.
Prof. Estévez is one of the founders of the Millennium Institute of Astrophysics (MAS), Chile, which was created in January 2014. He is currently leading the Astroinformatics/Astrostatistics group at MAS. He has been an Invited Researcher with the NTT Communication Science Laboratory, Kyoto, Japan; the Ecole Normale Supérieure, Lyon, France, and a Visiting Professor at the Pantheon-Sorbonne University, Paris, France, and the University of Tokyo, Tokyo, Japan.
Prof. Estévez is an IEEE Fellow. He served as President of the IEEE Computational Intelligence Society (CIS) for the term 2016-2017, Vice-president of Members Activities (2011-2014), CIS ADCOM Member-at-Large (2008-2010), CIS Distinguished Lecturer (2006-2011) and as an Associate Editor of the IEEE Transactions on Neural Networks (2007-2012). Prof. Estévez served as conference chair of the International Joint Conference on Neural Networks (IJCNN), held in July 2016, in Vancouver, Canada, and general co-chair of the 2018 IEEE World Congress on Computational Intelligence, IEEE WCCI 2018, held in Rio de Janeiro, Brazil, in July 2018. He is the recipient of the 2019 IEEE CIS Meritorious Service Award and the 2019 IEEE Region 9 Eminent Engineer Award.
Prof. Estévez has co-authored more than 150 articles in journals and conferences, including a recent paper published in Nature Astronomy. His current research interests include big data, deep learning, neural networks, self-organizing maps, data visualization, feature selection, information theoretic-learning, time series analysis, and advanced signal and image processing. One of his main topics of research is the application of machine learning and computational intelligence techniques to frontier research in astrophysics and biomedical engineering.
The dangers of post-hoc interpretability: unjustified counterfactual explanations
LIP6, Paris, France
AXA, Paris, France
Over the recent years, the field of XAI (eXplainable AI) has gained phenomenal interest. In particular, post-hoc interpretability approaches have been proven to be powerful tools to generate explanations for the predictions made by a trained black-box model. However, they create the risk of having explanations that are a result of some artifacts learned by the model instead of actual knowledge from the data. In this presentation, we identify several issues that can arise in this context and that may be harmful for interpretability. In particular, we focus on the case of counterfactual explanations and asks whether the generated instances can be justified, i.e. continuously connected to some ground-truth data. We evaluate the risk of generating unjustified counterfactual examples by investigating the local neighborhoods of instances whose predictions are to be explained and show that this risk is quite high for several datasets. Furthermore, we show that most state of the art approaches do not differentiate justified from unjustified counterfactual examples, leading to less useful explanations.
Thibault Laugel is defending his PhD in March 2020 from the Sorbonne University (Paris, France). He is working in the LIP6 laboratory under the supervision of Marie-Jeanne Lesot and Christophe Marsala, and in collaboration with the Research team of the AXA group, under the supervision of Marcin Detyniecki. His works, which focus on machine learning interpretability and adversarial learning, have been presented in several conferences specialized in these topics.
Learning to drive in Tokyo, driving in Paris
Professor at Sorbonne Université
LIP6, Paris, France
Matthieu.Cord [ at ] lip6.fr
Semantic segmentation is a key problem for many computer vision tasks. While approaches based on convolutional neural networks constantly break new records on different benchmarks, generalizing well to diverse testing environments remains a major challenge. In numerous real world applications, there is indeed a large gap between data distributions in train and test domains, which results in severe performance loss at run-time. In this talk, we address the task of unsupervised domain adaptation in semantic segmentation. We illustrate the strategies with two synthetic-2-real street-view image set-ups.
Matthieu Cord is a full professor at the Computer Science Laboratory (LIP6) of Sorbonne University since 2006. He is also working part-time at the valeo.ai research laboratory. He is a laureate of a chair of research and teaching in artificial intelligence from the national french government program on AI 2020 entitled VISA-DEEP: Towards visual reasoning in deep learning. He is an honorary member of the Institut Universitaire de France (2009) and served from 2015 to 2018 as AI expert at CNRS and ANR. His research expertise includes computer vision, machine learning and artificial intelligence. He is the author of more than 150 international scientific publications on visual information retrieval, pattern recognition using deep learning, and multimodal vision and language understanding.
An Overview of Evolutionary Multi-Objective Optimization
Carlos A. Coello Coello
IEEE CIS Vice President for Member Activities
Mexico City, México
ccoello [ at ] cs.cinvestav.mx
Multi-objective optimization refers to solving problems having two or more (often conflicting) objectives at the same time. Such problems are ill-defined and their solution is not a single solution but instead, a set of them, which represent the best possible trade-offs among the objectives.
Evolutionary algorithms are particularly suitable for solving multi-objective problems because they are population-based, and require little domain-specific information to conduct the search. Due to these advantages, the development of the so-called multi-objective evolutionary algorithms (MOEAs) has significantly increased in the last 15 years.
In this talk, we will provide a general overview of the field, including the main algorithms in current use as well as some of the many applications of them.
Carlos Artemio Coello Coello received a PhD in Computer Science from Tulane University (USA) in 1996. His research has mainly focused on the design of new multi-objective optimization algorithms based on bio-inspired metaheuristics, which is an area in which he has made pioneering contributions. He currently has over 500 publications which, according to Google Scholar, report over 48,500 citations (with an h-index of 86).
He is the recipient of the prestigious 2013 IEEE Kiyo Tomiyasu Award, "for pioneering contributions to single- and multiobjective optimization techniques using bioinspired metaheuristics" and of the 2016 The World Academy of Sciences (TWAS) Award in "Engineering Sciences". Since January 2011, he is an IEEE Fellow. He is also Associate Editor of several international journals including the IEEE Transactions on Evolutionary Computation, Evolutionary Computation and the IEEE Transactions on Emerging Topics in Computational Intelligence.
He is currently Vice President for Member Activities of the IEEE Computational Intelligence Society (CIS), and Full Professor with distinction at the Computer Science Department of CINVESTAV-IPN in Mexico City, Mexico.
Introduction to Quantum-Inspired Evolutionary Algorithms and their Application to Neural Architecture Search
IEEE CIS Vice President for Conferences
LIRA: Applied Computational Intelligence and Robotics Lab.
Pontifical Catholic University of Rio de Janeiro, Rio, Brazil
marley [ at ] ele.puc-rio.br
Quantum-inspired evolutionary algorithms can be considered a new class of evolutionary computation algorithms, inspired by quantum computing principles, that has been developed to achieve better performance in computationally intensive problems. In some applications, time taken to calculate the evaluation function may be critical. In those cases, reaching good solutions with the smallest possible number of evaluations is a very important factor. These quantum-inspired evolutionary models provide good (or optimal) solutions with a smaller number of evaluations, being adequate to solve optimization problems where the evaluation of each possible solution is computationally expensive. One important computationally expensive problem is the automatic configuration of a neural network, which, in the deep learning context is known as Neural Architecture Search.
In this talk we will present an overview of quantum-inspired evolutionary algorithms and their application to the evolution of different neural network models, focusing in the neural architecture search of Convolutional Neural Networks.
Marley Maria Bernardes Rebuzzi Vellasco received the BSc and MSc degrees in Electrical Engineering from the Pontifical Catholic University of Rio de Janeiro (PUC-Rio), Brazil, and the PhD degree in Computer Science from the University College London (UCL).
Dr. Vellasco is the founder and head of the Computational Intelligence and Robotics Laboratory (LIRA) of the Electrical Engineering Department at PUC-Rio.
She is the author of four books and more than 400 scientific papers in the area of soft computing and machine learning.
Her research interests are related to Computational Intelligence methods and applications, including Neural Networks, Fuzzy Logic, Hybrid Intelligent Systems (Neuro-Fuzzy, Neuro-Evolutionary and Fuzzy-Evolutionary models), applied to decision support systems, pattern classification, time-series forecasting, control, optimization and Data Mining.
Evolutionary Game AI
IEEE CIS Vice President for Education
School of Electronic Engineering and Computer Science
Queen Mary University of London, UK
simon.lucas [ at ] QMUL.AC.UK
Evolutionary algorithms can be applied in a number of useful ways to create intelligent game playing agents or to evolve or tune new games. In this talk I’ll explain how they can be used for game-playing agents, which falls into three distinct categories: rolling horizon evolution, policy evolution and hyper-parameter tuning. Rolling horizon evolution is a type of statistical forward planning algorithm that can be used as a direct alternative to Monte Carlo Tree Search, and often outperforms MCTS for video game AI. Note, in RL this process of searching for the best available action at each step is sometimes called Transient or Short-Term Learning; these methods require a fast and easily copied forward model of the game or simulation, but provide instant intelligent behaviour across a range of games. Policy evolution refers to directly evolving the parameters of a game-playing agent (for example, evolving the weights of a neural network). Learned policies can be combined with statistical forward planning algorithms to provide even higher standards of play. Finally, the hyper-parameters of these systems can be efficiently tuned using model-based evolutionary algorithms, such as the N-Tuple Bandit Evolutionary Algorithm. Model-based evolution often provide a significant boost in sample efficiency. In the talk I’ll demonstrate these algorithms in action on a number of games.
Simon Lucas is a professor of Artificial Intelligence and Head of the School of Electronic Engineering and Computer Science at Queen Mary University of London where he also heads the Game AI Research Group. He holds a PhD degree (1991) in Electronics and Computer Science from the University of Southampton. He is the founding Editor-in-Chief of the IEEE Transactions on Games and co-founded the IEEE Conference on Conference on Games. His research involves developing and applying computational intelligence techniques to build better game AI, use AI to design better games, provide deep insights into the nature of intelligence and work towards Artificial General Intelligence.
Reinforcement Learning of Strategies for General Game
Associate Professor at Sorbonne Université
LIP6, Paris, France
jean-noel.vittaut [ at ] lip6.fr
General Game Playing (GGP) is a domain of Artificial Intelligence (AI) aiming at developing autonomous agents which are able to play a large variety of games, called General Games. GGP is different from search algorithms because it allows to play specific games at a good level and opens the possibility to evaluate the efficiency of AI methods with no prior knowledge from experts.
An important aspect of our work lies in the utilization of an implicit game-tree representation as a set of logic rules, an explicit representation being too large to be stored in memory. In this context, we have proposed an efficient method of rule instantiation which allows the computation of a logic circuit making it possible to perform static and dynamic analyses of the game and simulate random matches more efficiently. By using Monte-Carlo Tree Search methods, we can learn strategies which we can start to use within time budgets compatible with human reflexion time.
Jean-Noël Vittaut has been an associate professor at the Computer Science Laboratory (LIP6) of Sorbonne University since 2019. He was previously a lecturer at Vincennes-Saint-Denis University and defended his PhD on General Game Playing (GGP) in 2017. His research focuses on machine learning applied to textual data and artificial intelligence for games. He has also developed a programme called LeJoueur which has won an international GGP competition.
Interval-valued non-additive image processing
Associate Professor at Montpellier University
LIRMM, Montpellier, France
strauss [ at ] lirmm.fr
Digital image processing consists of imitating continuous image processing such as contrast enhancement, restoration, rotation, denoising, etc. via an algorithm. These kinds of algorithms are constructed in such a way that the produced image resembles the 'digitalizing' of an image having undergone the imitated continuous process, as much as possible. For example, when a digital image is 'rotated', the idea is to create the digital image that one would have obtained by rotating the camera before shooting.
Most of these algorithms can be considered as linear aggregation processes, involving weighted sums. However, there is no single way to make an algorithm that mimics continuous image processing: each continuous process corresponds to an infinity of digital processes leading to an infinity of digital images. Each of these can be seen as the digitisation of the image having undergone the mimicked continuous process.
In recent work, we have shown that, by replacing the linear aggregation by an aggregation involving a fuzzy integral, it is possible to represent the convex set of this infinity of images. In this talk, we propose to present quickly the principle of digital image processing and its generalisation by fuzzy integrals.
Olivier Strauss received his Ph.D. in signal processing and systems in 1992 from the Montpellier University of Science, France. Since 1992 he has served as associate professor in the same University and researcher in the ICAR team (image and interaction) of the LIRMM (Laboratory of Informatics, Robotics and Microelectronic of Montpellier). His main research areas are image processing and vision. He has also a strong interest in sampling, epistemic statistics, imprecise probabilities and fuzzy sets.
The French Chapter on Computational Intelligence IEEE is happy to invite you to two events organised duting Prof. Dr. Rudolf Kruse's visit, subsidised under the IEEE Computational Intelligence Society Distinguished Lecturer Program and the IEEE France:
1. A lecture called « Decomposable Models: On Learning, Fusion and Revision » on March the 7th 2019 at 10h30
The lecture is open to all, see the abstract and biography below.
2. A masterclass named « Intelligent systems » with Prof. Dr. Rudolf Kruse, on March the 8th 2019 at 10h00
The event is open to all. PhD candidates can present their work, for 15 minutes, before the chat begins. All interested parties can make themselves known to Adrian Revault d'Allonnes (email@example.com) before March the 7th 2019, noon.
The events will take place at: LIP6, room 105 (1st floor), corridor 25-26, 4 place Jussieu, 75005 Paris.
Abstract: Decomposable Graphical Models are of high relevance for complex industrial applications. The Markov network approach is one of their most prominent representatives and an important tool to decompose uncertain knowledge in high dimensional domains. But also relational and possibilistic decompositions turn out to be useful to make reasoning in such domains feasible. Compared to conditioning a decomposable model on given evidence, the learning of the structure of the model from data as well as the fusion of several decomposable models is much more complicated. The important belief change operation revision has been almost entirely disregarded in the past, although the problem of inconsistencies is of utmost relevance for real world applications. In this talk these problems are addressed by presenting several successful complex industrial applications.
Short bio: Rudolf Kruse is Professor at the Faculty of Computer Science at University of Magdeburg in Germany. He obtained his Ph.D. and his Habilitation in Mathematics from the Technical University of Braunschweig in 1980 and 1984 respectively. Following a stay at the Fraunhofer Gesellschaft, he joined the Technical University of Braunschweig as a professor of computer science in 1986. Since 1996 he is a professor in the Computational Intelligence Group in Magdeburg. He has coauthored 15 monographs and 25 books as well as more than 350 peer-refereed scientific publications in various areas with 16000 citations. He is associate editor of several scientific journals. Rudolf Kruse is Fellow of the International Fuzzy Systems Association (IFSA), Fellow of the European Association for Artificial Intelligence (EURAI/ECCAI ), and Fellow of the Institute of Electrical and Electronics Engineers (IEEE). His group is successful in various industrial applications in cooperation with companies such as Volkswagen, SAP, Daimler, and British Telecom. His current main research interests include data science and intelligent systems.
The Computational Intelligence Chapter of IEEE France Section and the DAPA department of the LIP6 are proud to invite you to the following talk:
« Big Data Analytics using Deep Learning and Information Theoretical Learning: Applications to Astronomy »
given by Professeur Pablo A. Estévez, Department of Electrical Engineering, University of Chile, and Millennium Institute of Astrophysics, Chile, IEEE Computational Intelligence Society (2016-2017) President.
Date and time: 4 July 2017 at 10h
Venue: LIP6, salle 105 (1st floor), 25-26 corridor
4 place Jussieu, 75005 Paris.
Abstract: Astronomy is facing a paradigm shift caused by the exponential growth of the sample size, data complexity and data generation rates of new sky surveys. To cope with a change of paradigm to data-driven science new computational intelligence, machine learning and statistical approaches are needed. In this talk I will present two main applications. The first is to discriminate periodic versus non-periodic light curves, and then estimate the period of the periodic ones. Light curves are one-dimensional time series of the brightness of a star versus time. We have developed several methods based on the correntropy function (generalized correlation using information theoretical learning concepts), which outperforms conventional approaches. Results using 32.8 million light curves will be presented. Interestingly, some of these techniques can be applied to other problems such as sleep EEG analysis, and I will present preliminary results on this topic too.
The second application is the automated real-time transient detection in astronomical images. The aim is to achieve real-time detection of supernovae and other transients with the Dark Energy Camera. A novel transient detection pipeline was developed. We have been applying convolutional neural nets (deep learning) to discriminate between true transients and bogus transients, among other techniques, e.g non-negative matrix factorization combined with random forests. Results using 1.5 million images will be presented. The new pipeline was successfully tested online in February 2015 finding more than 100 supernovae in a few days of telescope observation.
Short bio: Pablo A. Estévez received his professional title in electrical engineering (EE) from Universidad de Chile, in 1981, and the M.Sc. and Dr.Eng. degrees from the University of Tokyo, Japan, in 1992 and 1995, respectively. He is a Full Professor with the Electrical Engineering Department, University of Chile, and former Chairman of the EE Department in the period 2006-2010.
Prof. Estévez is one of the founders of the Millennium Institute of Astrophysics (MAS), Chile, which was created in January 2014. He is currently leading the Astroinformatics/Astrostatistics group at MAS. He has been an Invited Researcher with the NTT Communication Science Laboratory, Kyoto, Japan; the Ecole Normale Supérieure, Lyon, France, and a Visiting Professor with the University of Tokyo.
Prof. Estévez is an IEEE Fellow. He is currently the President of the IEEE Computational Intelligence Society (CIS) for the term 2016-2017. He has served as IEEE CIS President-elect (2015), CIS Vice-president of Members Activities (2011-2014), CIS ADCOM Member-at-Large (2008-2010), CIS Distinguished Lecturer (2006-2011) and as an Associate Editor of the IEEE Transactions on Neural Networks (2007-2012).
Prof. Estévez served as conference chair of the International Joint Conference on Neural Networks (IJCNN), held in July 2016, in Vancouver, Canada, and general chair of the Workshop on Self-Organizing Maps (WSOM), held in December 2012, in Santiago, Chile. Currently he is serving as general co-chair of the 2018 IEEE World Congress on Computational Intelligence, WCCI 2018, to be held in Rio de Janeiro, Brazil, July 2018.
His current research interests include big data, deep learning, neural networks, self-organizing maps, data visualization, feature selection, information theoretic-learning, time series analysis, and advanced signal and image processing. One of his main topics of research is the application of computational intelligence techniques to astronomical datasets, and EEG signals.
Page web: link_to (https://www.researchgate.net/profil)
Talk by Professor Ronald R. Yager
IEEE Computational Intelligence Society Distinguished Lecturer
Multi-Criteria Decision Making and Uncertainty
L’IEEE France Section et l’IEEE Computational Intelligence Society ont le plaisir de vous inviter à la présentation du professeur Ronald R. Yager, directeur du Machine Intelligence Institute, Iona College, USA, reçu en tant que distinguished lecturer. La présentation portera sur la décision multicritère dans l'incertain, son résumé est disponible ci-dessous.
Date et horaire : 4 mai 2017 à 10H00
Lieu : LIP6, salle 105 (1er étage), couloir 25-26
4 place Jussieu, 75005 Paris.
Abstract: Multi-Criteria aggregation is a pervasive problem appearing in many technological domains. During this presentation, we shall discuss some issues related to this task. One issue is the modeling of multi-criteria decision functions and a related issue is the evaluation of these decision functions in the face of uncertain information. One case we shall consider is the evaluation of the OWA operator when the satisfaction to the individual criteria is expressed via a probability distribution. We shall also consider the case of interval criteria satisfactions. We shall look at the role of fuzzy measures in the modeling process. One issue that must be dealt with is the ordering of the complex uncertain criteria satisfactions that is required to use the Choquet integral in the criteria aggregation.
Bio: Ronald R. Yager is Director of the Machine Intelligence Institute and Professor of Information Systems at Iona College. He is editor-in-chief of the International Journal of Intelligent Systems. He has published over 500 papers and edited over 30 books in areas related to fuzzy sets, human behavioral modeling, decision-making under uncertainty and the fusion of information. He is among the world’s most highly cited researchers with over 57,000 citations in Google Scholar. He was the 2016 recipient of the IEEE Frank Rosenblatt Award, the most prestigious honor given out by the IEEE Computational Intelligence Society. He was the recipient of the IEEE Computational Intelligence Society Pioneer award in Fuzzy Systems. He received the special honorary medal of the 50-th Anniversary of the Polish Academy of Sciences. He received the Lifetime Outstanding Achievement Award from the International Fuzzy Systems Association. He received honorary doctorate degrees, honoris causa, from the Azerbaijan Technical University and the State University of Information Technologies, Sofia Bulgaria. Dr. Yager is a fellow of the IEEE, the New York Academy of Sciences and the Fuzzy Systems Association. He has served at the National Science Foundation as program director in the Information Sciences program. He was a NASA/Stanford visiting fellow and a research associate at the University of California, Berkeley. He has been a lecturer at NATO Advanced Study Institutes. He was a visiting distinguished scientist at King Saud University, Riyadh Saudi Arabia. He was an honorary professor at Aalborg University in Denmark. He received his undergraduate degree from the City College of New York and his Ph. D. from the Polytechnic Institute New York University. He recently edited a volume entitled Intelligent Methods for Cyber Warfare.
I am pleased to announce that the IEEE Computational Intelligence Society President will visit us on Thursday, 22nd of October at 10h, in the LIP6, to give the following talk:
Ensemble Approaches in Learning
by Xin Yao, Professor at the University of Birmingham (UK), President of the IEEE Computational Intelligence Society
Venue: room 105, corridor 25-26, 4 place Jussieu, 75005 Paris
Designing a monolithic system for a large and complex learning task is hard. Divide-and-conquer is a common strategy in tackling such large and complex problems. Ensembles can be regarded an automatic approach towards automatic divide-and-conquer. Many ensemble methods, including boosting, bagging,
negative correlation, etc., have been used in machine learning and data mining for many years. This talk will describe three examples of ensemble methods,
i.e., multi-objective learning, online learning with concept drift, and multi-class imbalance learning. Given the important role of diversity in ensemble methods, some discussions and analyses will be given to gain a better understanding of how and when diversity may help ensemble learning.
Some materials used in the talk are based on the following papers:
A Chandra and X. Yao, ``Ensemble learning using multi-objective evolutionary
algorithms,'' Journal of Mathematical Modelling and Algorithms, 5(4):417-445,
L. L. Minku and X. Yao, "DDD: A New Ensemble Approach For Dealing With Concept
Drift,'' IEEE Transactions on Knowledge and Data Engineering, 24(4):619-633,
S. Wang and X. Yao, ``Multi-Class Imbalance Problems: Analysis and Potential
Solutions,'' IEEE Transactions on Systems, Man and Cybernetics, Part B,
42(4):1119-1130, August 2012.
Xin Yao is a Chair (Professor) of Computer Science and the Director of CERCIA (Centre of Excellence for Research in Computational Intelligence and
Applications) at the University of Birmingham, UK. He is an IEEE Fellow and the President (2014-15) of IEEE Computational Intelligence Society (CIS). His
work won the 2001 IEEE Donald G. Fink Prize Paper Award, 2010 and 2015 IEEE Transactions on Evolutionary Computation Outstanding Paper Awards, 2010 BT Gordon Radley Award for Best Author of Innovation (Finalist), 2011 IEEE Transactions on Neural Networks Outstanding Paper Award, and many other best
paper awards. He won the prestigious Royal Society Wolfson Research Merit Award in 2012 and the 2013 IEEE CIS Evolutionary Computation Pioneer Award.
His major research interests include evolutionary computation, ensemble learning, and their applications, especially in software engineering.
More information on Xin Yao: http://www.cs.bham.ac.uk/~xin/
IEEE France Section Computational Intelligence Chapter President
We are pleased to announce that, on the 28th of May 2015, Prof. Jim Keller, of the University of Missouri, USA, will give a talk entitled "Does it all add up? A study of fuzzy protoform linguistic summarization of time series", sponsored by the Computational Intelligence Society under its Distinguished Lecturer Program. The abstract for this talk is:
Producing linguistic summaries of large databases or temporal sequences of measurements is an endeavor that is receiving increased attention. These summaries can be used in a continuous monitoring situation, like eldercare, where it is important to ascertain if the current summaries represent an abnormal condition. Primarily a human, such as a care giver in the eldercare
example, is the recipient of the set of summaries describing a time range, for example, last night’s activities. However, as the number of sensors and monitored conditions grow, sorting through a fairly large number of summaries can be a burden for the person, i.e., the summaries stop being information and become yet one more pile of data. It is therefore necessary to automatically process sets of summaries to condense the data into more manageable chunks.
The first step towards automatically comparing sets of digests is to determine similarity. For fuzzy protoform based summaries, we developed a natural similarity and proved that the associated dissimilarity is a metric over the space of protoforms. Utilizing that distance measure, we defined and examined several fuzzy set methods to compute dissimilarity between sets of summaries, and most recently utilized these measures to define prototypical behavior over a large number of normal time periods.
In this talk, I will cover the definition of fuzzy protoforms, define our (dis)similarity, outline the proof that it is a metric, discuss the fuzzy aggregation methods for sets of summaries, and show how prototypes are formed and can used to detect abnormal nights. The talk will be loaded with actual examples from our eldercare research. There is much work to be done and hopefully, more questions than answers will result from the discussion.
An up to date list of our seminars can be found here and a list of coming ev\
ents is here.