FUZZ-IEEE has a long tradition of inviting illustrious speakers, which have strongly contributed with their research to the area of fuzzy logic theory and applications. Also for the 2017 edition, plenary talks are chosen with an eye to highlighting state-of-the-art hot topics in fuzzy logic area, providing retrospectives of technology advances, and highlighting the high applicability of fuzzy sets to different and heterogenous application domains. The FUZZ-IEEE 2017 Program Committee is pleased to announce that the following plenary presentations (in alphabetical order with respect to the last name of speakers) will be held in the beautiful scenario of Naples.
Computer vision is a well-known area where computational intelligence has made a significant impact. In general, the field is diverse and objectives range from filtering to object detection, image understanding and linguistic summarization/description, to name a few. As simple as it may sound, we have been trying to make a computer “describe what it saw” since the 1960s. In an attempt to achieve this goal, researchers have looked to data/information fusion. However, most classical aggregation strategies are additive and assume independence among inputs. On the other hand, fuzzy measure theory provides a powerful parametric way to specify or learn input interactions (when/if available). More importantly, the fuzzy integral utilizes the fuzzy measure to achieve nonlinear aggregation. In this talk, I will discuss the role of nonlinear aggregation via fuzzy integrals at the levels of signal, spectrum, feature, and decision-level fusion. In particular, I highlight recently established extensions of fuzzy integrals designed to address key challenges in computer vision. These extensions focus on spatial and/or distribution level uncertainty and they are embedded into pattern recognition or automated decision making via multiple kernel learning and/or fuzzy logic. Applications are discussed for multi-sensor humanitarian demining, hyperspectral image analysis and remote sensing.
Craniofacial superimposition is a skeleton-based, forensic identification technique that can provide evidence to support that some human skeletal remains belong or not to a missing person. The process aims to overlay a skull with some ante-mortem images of a candidate in order to determine if they correspond to the same person. Although craniofacial superimposition has been in use for over a century, there is not a common methodology accepted worldwide. Instead, each forensic anthropologist applies a specific approach considering her expert knowledge and the available technologies. Hence, there is a strong interest in designing systematic and automatic methods to support the forensic anthropologist to apply craniofacial superimposition, avoiding the use of subjective, error-prone, and time-consuming manual procedures. The use of computational intelligence is a natural way to achieve this aim. In particular, evolutionary algorithms and fuzzy sets can properly be used to automate the procedure while handling the underlying uncertainty. This talk is devoted to present an intelligent system for craniofacial identification developed in collaboration with the University of Granada’s Physical Anthropology Lab within a ten year long research project. Our system is composed of a three-stage procedure involving the automatic reconstruction of 3D models of human skulls using evolutionary algorithms and image registration methods, the obtaining of 3D skull model – 2D face photograph overlays based on evolutionary algorithms and fuzzy sets, and the determination of a degree of support for the assertion that that the skull and the ante-mortem image belong to the same person by a decision support system based on fuzzy aggregations and computer vision techniques. The resulting system is protected by an international patent and is currently under commercialization in Mexico. The results obtained in several real-world cases solved by the Physical Anthropology Lab in cooperation with the Spanish Scientific Police will be reported.
Type-2 fuzzy sets and systems, including both interval and general type-2 sets, are now firmly established as tools for the fuzzy researcher that may be deployed on a wide range of applications and in a wide set of contexts. However, in many situations the output of type-2 systems are type-reduced and then defuzzified to an interval centroid, which are then often even simply averaged to obtain a single crisp output. Many successful applications of type-2 have been in control contexts, often focussing on reducing the RMSE. This is not taking full advantage of the extra modelling capabilities inherent in type-2 fuzzy sets. In this talk, I will present some of the current research being carried out within the LUCID group at Nottingham, and wider, into type-2 for modelling human reasoning, including approaches and methodologies which make more use of type-2 capabilities.
Brain-Computer Interface (BCI) enhances the capability of a human brain in communicating and interacting with the environment directly. BCI plays an important role in natural cognition, which concerns the studies of brain and behavior at work for enhancing or restoring cognitive functions. Many people may benefit from BCI, which facilitates continuous monitoring of fluctuations in cognitive states under monotonous conditions in workplace or at home. People who suffer from episodic or progressive cognitive impairments in daily life can also benefit from BCI. In this talk, I will first introduce the current status of BCI and its major obstacles: lack of wearable EEG devices, various forms of noise contamination, user/circadian variability, and lack of suitable adaptive cognitive modeling. I will then introduce some methodologies to overcome these obstacles, including discovering the fundamental physiological changes of human cognitive functions at work and then utilizing these main bio-findings and computational intelligence (CI) techniques to monitor, maintain, or track human cognitive states and operating performance. In the second part of my presentation, I will introduce an innovative BCI-inspired research domain called Cyber-Brain-Physical Systems. Some future research directions in this domain will be explored and discussed, including BCI-embedded wearable computing, BCI-based neuro-prosthesis and assistive devices, wearable cognitive robots, and BCI-empowered training. The potential real-life applications of BCI on various aspects of training/education, healthcare, rehabilitation, and medical treatment will also be introduced and discussed.
Various processes related to the task of Information Retrieval (IR) can be interpreted as Multi Criteria Decision Making activities. The same applies to some tasks related to the analysis of user generated content in Social Media (like for example the assessement of veracity of online reviews). What is particularly interesting in this interpretation is the role of aggregation operators, which, for a given alternative, reduce the performace-scores of the considered criteria into a global performace-score of the alternative. In fact, depending on the selected aggregation strategy, different behaviors can be modelled for the considered process, corresponding to distinct predictive models. These behaviors can be more intuitively captured by guiding the aggregation by means of fuzzy quantifiers (quantifiers guided aggregation). Formally, this can be achieved by employing fuzzy integrals and quantifier guided OWA aggregation. As an example, in Information Retrieval the assessment of the relevance a document (an alternative) to a query can be seen as the process of evaluating the performance of several relevance dimensions (criteria) like topicality, novelty, recency, etc. In relation to the analysis of user generated content, an example is offered by the assessment of the veracity of an online review (alternative), which is based on several features (criteria). In this lecture the impact of quantifier guided aggregation (and of aggregation in general) will be shown in both contexts of IR and of the assessment of veracity of user generated contents. It will be also shown that this quantifier guided aggregation offers an interesting alternative to the application of machine learning techniques (in particular classifiers).
When dealing with navigation/control of (semi-) autonomous robotic vehicles in obstacle filled dynamic environments, Fuzzy Logic offers a reliable and viable alternative to conventional controller design and analytic techniques, as it is capable of handling environment uncertainty that is difficult if not impossible to model, as well as system modeling uncertainties without affecting system robustness nor adversely impacting performance.
This talk presents a generalized Fuzzy Logic based hierarchical architecture and framework along with its application specific modifications for aerial, aquatic and terrestrial robotic vehicle sensor-based autonomous navigation and control. For such applications, a mathematical model of the dynamics of the vehicle is not needed during the design process of the motion controller; however, the problem-specific heuristic control knowledge is needed for the inference engine design. From the practical and implementation point of view, it is shown that Fuzzy Logic is the most appropriate modeling tool to represent imprecision and uncertainty of sensor readings, and for hardware implementation of fuzzy controllers in real-time due to low computation time.
Experimental and simulation studies and results validate and support implemented techniques and approaches to ground, aerial and underwater vehicles, followed by a comparative study of classical and soft computing based controllers designed to control small unmanned helicopters.