IISA2016 | Deep Learning vs. Wise Learning: A Critical and Challenging Overview
singular,single,single-post,postid-8056,single-format-standard,ajax_fade,page_not_loaded,,qode-child-theme-ver-1.0.0,qode-theme-ver-3.1,wpb-js-composer js-comp-ver-4.11.1,vc_responsive

Deep Learning vs. Wise Learning: A Critical and Challenging Overview

Deep Learning vs. Wise Learning: A Critical and Challenging Overview

  |   Invited Keynote Speakers   |   No comment

Peter P. Groumpos

Professor and Director of the Laboratory for Automation and Robotics
Department of Electrical and Computer Engineering
University of Patras, Greece


In this keynote paper the most important scientific challenge of knowledge learning is reviewed thought two different approaches: DEep LEarning (DE.LE) and WIse LEarning (WI.LE).

Learning is the most important thing that living creatures do. As far as any living creature is concerned, any action that does not involve learning is pretty much a waste of time. This is especially so for a human one. An organism cannot properly animate itself without first learning how to. Humans, before they can satisfy their own needs, first have to learn how to do it.

DEep LEarning (DE.LE) is a new branch of machine learning based on a set of algorithms that attempt to model high-level abstractions in data by using multiple processing layers, with complex structures or otherwise, composed of multiple non-linear transformations. Research in this area attempts to make better representations and create models to learn these representations from large-scale unlabeled data. Some of the representations are inspired by advances in neuroscience and are loosely based on interpretation of information processing and communication patterns in a nervous system, such as neural coding which attempts to define a relationship between various stimuli and associated neuronal responses in the brain.

In this keynote paper DL is extensively reviewed. Various deep learning architectures such as deep neural networks, convolutional deep neural networks, deep belief networks, recurrent neural networks among other ones are reviewed. The way that different DL algorithms have been applied to a number of fields like computer vision, automatic speech recognition, natural language processing, audio recognition, expert systems and bioinformatics where they have been shown to produce state-of-the-art results on various tasks are presented and analyzed. Specific examples are presented.

WIse LEarning (WI.LE) is a simple new emerging mathematical approach for modeling Complex Dynamic Systems (CDS). It is new but very challenging. It is exploring and combining theories of fuzzy logic, neural networks, artificial intelligence, intelligent control, data trees, fuzzy cognitive maps, Hebbian learning and other advanced state space methods in a new integrated approach. The new main ingredient of this new approach is that “wise” constraints must be taken into considerations. The basis of the new approach is Fuzzy Cognitive Maps which explores simultaneously: Knowledge Base Systems (KBS), computational complexity, Smart Logistic Systems (SLS), Cloud Computing and Internet of Things, multiyear historical data, Multiagent Systems using many experts and discrete state space techniques. A critical comparison between Deep Learning (DE.LE) and Wise Learning (WI.SE) approaches is presented and future research directions in these two different approaches are presented.

Short Bio

Professor P. Groumpos received his Ph.D. from the University of Buffalo of SUNY, in 1978, Department of Electrical Engineering and has been engaged in teaching, research and collaboration with industry since then both in USA (1979-1989) and in Greece since his return from USA to the University of Patras in 1990. He is full professor at the Dept. of EE&CS of the University of Patras. In USA he developed and directed an Energy Research Center at Cleveland State University (1981-1989) while in Greece he developed the Laboratory for Automation and Robotics (LAR) in 1991 and he is its Director since then. His research interests are in the areas of modeling complex dynamic systems, energy, environment, renewable energy sources, Robotics, Fuzzy Cognitive maps (FCM), techno and economics for intelligent buildings and smart grids, optimization of electric power systems and Intelligent Control. He has used extensively Fuzzy Cognitive Maps (FCMs) in many scientific fields with very useful research results such as in: health, energy, environment, communications, agriculture, transportation, Business and Economics Hybrid Energy Systems and reliability analysis of engineering systems.

He is the author of more than 300 scientific publications (Books, invited book chapters, articles, papers, monographs and technical reports). He was the President of Patras Science Park (2004-2010). He has been as principal Investigator in more than 20 Projects as well as partner in another 40 projects.
He is an Invited Academician honorary member of the Russian Council on Mechatronics and Robotics since 2002 and Guest honorary Professor of the East China University of Science and Technology since 2013. He has been the Greek National representative to a number of EU Program Management committees (Esprit, ICT, IMS, and INCO). Professor Groumpos has given more than 15 Plenary and/or Keynote Presentations in International Conferences the last 10 years.