| |
Last updated on October 16, 2022. This conference program is tentative and subject to change
Technical Program for Sunday October 9, 2022
To show or hide the keywords and abstract (text summary) of a paper (if available), click on the paper title
Open all abstracts
Close all abstracts
Presentation  In person  On-line  No presentation  No information
|
Su-S1-TU1 Tutorial Session, NADIR |
Add to My Program |
Model-Based Systems Engineering of Information Fusion Systems and
Evaluation with Machine Learning and Statistical Methods (Part 1) |
|
|
Chair: Ali K Raz, Ali | George Mason University |
|
14:00-16:00, Paper Su-S1-TU1.1 | Add to My Program |
Model-Based Systems Engineering of Information Fusion Systems and Evaluation with Machine Learning and Statistical Methods |
|
Ali K Raz, Ali | George Mason University |
Keywords: System Architecture, Large-Scale System of Systems, Decision Support Systems
Abstract: Information fusion (IF) systems find their application in multiple domains from defense applications to self-driving cars, and autonomous systems. Irrespective of the application domain, an IF system’s objective is to produce optimal state/situation estimates from various sources that are supportive of typically mixed-initiative decision-making which leads to an action. These three processes of fusion, sensemaking, and decision-making have critical interdependencies that are often overlooked in an IF system design. Engineering the IF system requires a holistic, systemic perspective that includes evaluation of a multitude of interacting design variables which span the fusion, sensemaking and decision-making aspects. This tutorial addresses this problem from a Systems Engineering approach and evaluation methodology. In this tutorial, first the interdependencies between fusion, sensemaking, and decision-making are introduced, followed by a development of a domain-agnostic framework which provides holistic design and performance evaluation of an IF system. A system development framework is presented that leverages Systems Engineering principles and Model-based Systems Engineering (MBSE) techniques for practical system development. On the evaluation side, this tutorial pairs the MBSE approach with statistical methods and machine learning to provide a holistic and integrated evaluation of fusion, sensemaking, and decision-making. Theoretical foundations for performing design and analysis of experiments, followed by a hands-on IF system application example is presented. A refresher on Monte-Carlo simulations and hypothesis testing will also be provided.
|
|
Su-S1-TU2 Tutorial Session, AQUARIUS |
Add to My Program |
A Brief Introduction on Fractional Fuzzy Inference Systems: The New
Generation of Fuzzy Inference Systems |
|
|
Chair: Mazandarani, Mehran | Shenzhen University |
|
14:00-16:00, Paper Su-S1-TU2.1 | Add to My Program |
A Brief Introduction on Fractional Fuzzy Inference Systems: The New Generation of Fuzzy Inference Systems |
|
Mazandarani, Mehran | Shenzhen University |
Keywords: Control of Uncertain Systems, Decision Support Systems, Modeling of Autonomous Systems
Abstract: The year 2020 came with a new generation of fuzzy inference systems (FISs) called fractional fuzzy inference systems (FFISs). FFISs add to the branch of the calculus of fuzzy rules and fuzzy graphs some new concepts of which are namely: fractional membership functions, fractional compositional rule of inference, fracture index, left and right orders, and so forth. This tutorial sheds light on the reasons for a transition from typical fuzzy inference systems (e.g., Mamdani’s FIS) to their corresponding FFISs. The impetus for the transition derives from the fact that both the generality of an FIS and its applicability to real-world problems are substantially enhanced by replacing the concept of the typical inference mechanism with that of a fractional inference mechanism. The conceptual impressions related to this new generation of FISs are elaborated in the tutorial, and the advantages of FFISs over typical FISs are also explained. In addition, potential applications of FFISs are highlighted, focusing on expert and knowledge-based systems and computational intelligence. Moreover, it has been proved that independent of the problem under consideration, typical FISs never lead to more satisfactory results than those obtained by FFISs. Thus, applying FFISs in the place of typical FISs brings us higher performance, more accuracy, less error, more efficiency, higher quality, more accurate prediction, and eventually increases the degree of computational intelligence.
|
|
Su-S1-TU3 Tutorial Session, TAURUS |
Add to My Program |
Introduction to Hardware-Aware Neural Architecture Search |
|
|
Chair: Sekanina, Lukas | Brno University of Technology |
|
14:00-16:00, Paper Su-S1-TU3.1 | Add to My Program |
Introduction to Hardware-Aware Neural Architecture Search |
|
Sekanina, Lukas | Brno University of Technology |
Keywords: Neural Networks and their Applications, Machine Learning, Computational Intelligence
Abstract: As deep neural networks (DNNs) can have complex architectures with millions of trainable parameters, their design and training are difficult even for highly qualified experts. Neural architecture search (NAS) methods have been developed to automate the entire design process. With the aim of reaching desired latency and providing high energy efficiency, specialized hardware accelerators for DNN inference were developed for cutting-edge applications. In this direction, hardware-aware NAS methods were adopted to design DNN architecture (and weights) optimally for a given hardware platform. In this tutorial, we survey the critical elements of NAS methods that – to various extents – consider hardware implementation of the resulting DNNs. We will classify these methods into three major classes: single-objective NAS (no hardware is considered), hardware-aware NAS (DNN is optimized for a particular hardware platform), and NAS with hardware co-optimization (hardware is directly co-optimized with DNN as a part of NAS). We emphasize the multi-objective design approach that must be adopted in NAS and focus on co-design algorithms developed for concurrent optimization of DNN architectures and hardware platforms. As most research in this area deals with NAS for image classification using convolutional neural networks, our case studies will be devoted to this application.
|
|
Su-S1-TU4 Tutorial Session, LEO |
Add to My Program |
Implementing Evolutionary Optimization on Actual Quantum Processors |
|
|
Chair: Acampora, Giovanni | University of Naples Federico II |
|
14:00-16:00, Paper Su-S1-TU4.1 | Add to My Program |
Implementing Evolutionary Optimization on Actual Quantum Processors |
|
Acampora, Giovanni | University of Naples Federico II |
Vitiello, Autilia | University of Naples Federico II |
Keywords: Quantum Machine Learning, Evolutionary Computation, Computational Intelligence
Abstract: Quantum computing is a new paradigm that uses quantum phenomena such as superposition, entanglement, and interference to perform computation in a more efficient way than classical computers. Currently, quantum computing is showing its benefits in artificial intelligence, with a focus on machine learning domain where quantum algorithms are used to support the training of complex models for classification, regression, and clustering. In our vision, evolutionary computation is another area that could greatly benefit from the quantum revolution, due to the intrinsically parallel and distributed nature of evolutionary algorithms. This tutorial provides auditors with basic knowledge for implementing and testing a hybrid classical/quantum genetic algorithm, namely HQGA, where some key evolutionary steps, such as chromosome creation and storage, crossover, mutation, and selection are performed on an actual quantum computer, whereas other computations such as fitness function evaluations are performed on a classical computer. Auditors will be supported with a set of Python notebooks where actual quantum computers belonging to the IBM Q Experience initiative are used to run HQGA on a set of well-known benchmark functions, to put their skills in programming a quantum computer with a genetic algorithm into practice.
|
|
Su-S1-TU5 Tutorial Session, VIRGO |
Add to My Program |
Interaction-Centered Design for AI-Enabled Socio-Technical Systems |
|
|
Chair: Hou, Ming | Department of National Defence, Canada |
|
14:00-16:00, Paper Su-S1-TU5.1 | Add to My Program |
Interaction-Centered Design for AI-Enabled Socio-Technical Systems |
|
Hou, Ming | Department of National Defence, Canada |
Keywords: Design Methods, Intelligence Interaction, Human-Machine Cooperation and Systems
Abstract: With collective human-machine intelligence, the human-machine symbiosis (HMS) technologies are prevalent in society today and capable of solving complex problems. However, the trend raises important questions about the complications, liabilities, risks, and trust associated with increasing intelligence and adaptivity in these socio-technical systems given both humans and machines have limitations. It is even more challenging when we are facing insufficient data, indeterministic conditions, and inexhaustive solutions for uncertain actions. Recent accidents to the Boing 737 Max passengers ring the alarm about the importance of the appropriate design methodologies for safety-critical socio-technical systems. To provide guidance to understand and mitigate the potential risks associated with employing these HMS technologies, the evolution of design strategy and methodology for HMS technologies will be reviewed first. An interaction-centered design (ICD) framework, an associated set of methodologies and roadmap, and a related trust model called IMPACTS will also be introduced as an enduring strategy and appropriate solution. Validation studies on the utility and effectiveness of the ICD approach through real-world military technology evaluation activities will be presented. Challenges in integrating the ICD approach into systems engineering and validation processes will then be discussed for the future directions in R&D and exploitation of HMS technologies.
|
|
Su-S1-WS1 Workshop Session, KEPLER |
Add to My Program |
Human-Technology Embodied Integration Strategies for Patients’ Engagement |
|
|
Chair: Barresi, Giacinto | Istituto Italiano Di Tecnologia |
|
14:00-16:00, Paper Su-S1-WS1.1 | Add to My Program |
Human-Technology Embodied Integration Strategies for Patients’ Engagement |
|
Barresi, Giacinto | Istituto Italiano Di Tecnologia |
Marta, Matamala-Gomez | University of Barcelona |
Keywords: Human-Machine Cooperation and Systems, Virtual and Augmented Reality Systems, Human Factors
Abstract: Engaging patients in clinical routines is a crucial step to enhance their adherence to the treatment and, consequently, the therapeutic outcome. In particular, solutions for patients’ engagement in technological settings can be promoted through different user-centered design approaches. For instance, patient’s engagement can be considered as a result of human-technology integration, through which is possible to induce processes like the sense of embodiment (the feeling of a non-bodily object as a body part and a source of bodily sensations). Summoning such phenomena implies that the design of digital and mechatronic solutions for applications like pain treatment, upper limb rehabilitation or prosthetic training can reach an optimal matching between patient and technology. Demonstrations of this can be observed across wearable robotic systems, bionic prostheses, virtual and augmented settings, and other technological systems, possibly intertwining their peculiarities for promoting the patient’s engagement to a repetitive exercise or the acceptance of an artificial limb. This workshop aims at providing deep knowledge about the impact of inducing the sense of embodiment mediated by different technological solutions, alongside heterogeneous strategies for human-technology integration in clinical (and experimental) settings with the goal of highlighting how treatments can be enhanced through research in user-centered interaction design.
|
|
Su-S1-WS3 Workshop Session, TYCHO |
Add to My Program |
NeuroDesign in Human-Robot Interaction: The Making of Engaging HRI
Technology Your Brain Can’t Resist (Part 1) |
|
|
Chair: Wang, Ker-Jiun | University of Pittsburgh |
|
14:00-16:00, Paper Su-S1-WS3.1 | Add to My Program |
NeuroDesign in Human-Robot Interaction: The Making of Engaging HRI Technology Your Brain Can’t Resist |
|
Wang, Ker-Jiun | University of Pittsburgh |
Mao, Zhi-Hong | University of Pittsburgh |
Sugaya, Midori | Shibaura Institute of Technology |
Keywords: Human-Machine Interface, Human-Machine Cooperation and Systems, Brain-based Information Communications
Abstract: The IEEE SMC 2022 workshop on “NeuroDesign in Human- Robot Interaction: The making of engaging HRI technology your brain can’t resist” provides a mind-blowing, all-in-one forum to address the issue of how to transform HRI-related lab research into practical consumer product by using neuroscience and psycho-behavioral principles and approaches, such that the designed human-robot interaction strategies will be so attractive that complies to our brain’s natural perception and cognitive processes, and therefore overcome the barriers of social adoptions of using robots in every corner of our society. As a sequel of the three successful IEEE RO-MAN workshops in 2018, 2019 and 2020, this episode will provide a series of lectures from the latest advancement of HRI and Neuro/Brain-technology research to the design thinking and empathy-centric product development, as well as the commercialization strategies specifically tailored to the HRI & Neurotech products to solve the real-world problems. On top of the informative lectures/tutorials from both academia and industry experts, we also have Innovation Showcase competition, where the students, research labs, and startups could demonstrate their latest Brain-Machine Interface (BMI) and its applications of human-robot interactions to win grand prizes in our workshop. In the meanwhile, all the participants can have a perfect networking venue to brainstorm how to design intuitive brain-machine interaction dynamics for practicality and seek collaboration opportunities to solve the real-world social challenges. The purpose of this workshop is aiming at fulfilling the unmet need for lab research, to merge the scientific/theoretical findings in HRI, neuroscience, AI research communities and the real-world problems, targeting at end-users, focusing on the design & engineering journey of taking practical HRI and neuroscience solutions from labs to our daily lives, where everyone could enjoy using promising technologies avidly to improve our overall well-being.
|
|
Su-S2-TU6 Tutorial Session, AQUARIUS |
Add to My Program |
Making Model-Based System Design Aware of Design Requirements: Challenges
in Using and Extending SysML |
|
|
Chair: Nikolaidou, Mara | Harokopio University of Athens |
|
16:30-18:30, Paper Su-S2-TU6.1 | Add to My Program |
Making Model-Based System Design Aware of Design Requirements: Challenges in Using and Extending SysML |
|
Nikolaidou, Mara | Harokopio University of Athens |
Keywords: Large-Scale System of Systems, System Architecture, System Modeling and Control
Abstract: Model-based system design is served by a single, multi-level system model supporting all design activities, in different levels of detail. SysML, the OMG standard introduced in 2007, provides the means for defining such models to system designers. Over this period, SysML, currently in its 7th edition, has been adopted by both the academia and the industry and is moving towards its second version. However, engineers are still skeptical towards SysML. Studies revile that some of the main reasons identified by engineers are a) there are no user-oriented methodologies to facilitate its use, as many of them claim “it is still to expensive and unmanageable”, b) there is a lack of automation and integration with external tools and c) domain-specific profiles are not widely available. To promote SysML usage in model-based design, in this tutorial we discuss our efforts to establish a model-based approach for system design focusing on the exploitation and verification of design requirements expressed in a quantitative fashion. SysML extensions to handle and compute complex non-functional requirements, such performance, quality or cost and verify their support by system components. To this end, a consistent fashion to describe requirements and their verification formulas shall be presented. The presented framework provides for the automated integration of external tools for simulation, model analysis and decision-making during system design. All information created by external tools during system design is integrated within SysML system model, while the system designer only interacts with SysML to perform all engineering activities. External tool characteristics (for example simulation language) as well as detailed, low level information (for example simulation results) are hidden from the designer. T Participants will be provided with a brief overview of SysML and the proposed methodology, an introduction to the usage of external tools and the necessary SysML extensions. Furthermore, two domain-specific profiles, based on the proposed concepts, will be analytically presented. One related to the operation of railway transportation system and one to the design of Cyber-physical Human Systems (CPHS).
|
|
Su-S2-TU7 Tutorial Session, TAURUS |
Add to My Program |
Privacy-Preserving Federated Learning |
|
|
Chair: Tianfield, Hua | Glasgow Caledonian University |
|
16:30-18:30, Paper Su-S2-TU7.1 | Add to My Program |
Privacy-Preserving Federated Learning |
|
Tianfield, Hua | Glasgow Caledonian University |
Keywords: Machine Learning, Deep Learning
Abstract: Common privacy-enhancing technologies (e.g., differential privacy, homomorphic encryption, secure multi-party computation, etc.) are needed to provide an effective paradigm of secure collaboration. Federated learning reduces the risk of data security and privacy, providing a new solution to cross-organisational data-driven collaborations. However, simple architecture of federated learning still suffers from some security / privacy problems. This tutorial will provide an overview on the security / privacy issues of federated learning. In particular, the tutorial will look in to the methods of enhancing the privacy of federated learning, and will illustrate the blockchain based federated learning architecture. At the end, the tutorial will give a case study of health application to illustrate the concepts and frameworks.
|
|
Su-S2-TU8 Tutorial Session, LEO |
Add to My Program |
Validating Smart Energy Systems |
|
|
Chair: Strasser, Thomas | AIT Austrian Institute of Technology |
|
16:30-18:30, Paper Su-S2-TU8.1 | Add to My Program |
Validating Smart Energy Systems |
|
Strasser, Thomas | AIT Austrian Institute of Technology |
Keywords: Intelligent Power Grid, Infrastructure Systems and Services, Smart Buildings, Smart Cities and Infrastructures
Abstract: A driving force for the realization of a sustainable energy supply is the integration of renewable energy resources. Due to their stochastic generation behaviour, energy utilities are confronted with a more complex operation of the underlying power grids. Additionally, due to technological developments, controllable loads, integration with other energy sources, changing regulatory rules, and market liberalization, the system’s operation needs adaptation. Proper operational concepts and intelligent automation provide the basis to turn the existing power system into an intelligent entity, a smart grid. While reaping the benefits that come along with those intelligent behaviours, it is expected that system-level developments and testing will play a significantly larger role in realizing future solutions and technologies. Proper validation approaches, concepts, and tools are partly missing until now. This tutorial aims to tackle the above-mentioned requirements by introducing validation methods and tools for validating smart energy systems which are currently being developed in the European project ERIGrid 2.0.
|
|
Su-S2-TU9 Tutorial Session, VIRGO |
Add to My Program |
Computational Social Simulation Using E-CARGO |
|
|
Chair: Zhu, Haibin | Nipissing University |
|
16:30-18:30, Paper Su-S2-TU9.1 | Add to My Program |
Computational Social Simulation Using E-CARGO |
|
Zhu, Haibin | Nipissing University |
Keywords: Distributed Intelligent Systems, Decision Support Systems, System Architecture
Abstract: Humans are social beings and people cannot live alone. Computational social simulation is a way to reproduce a real-world society and study the behavior of people in that society using computer-based systems. Computational social simulation is a long-term, cutting-edge topic in the interdisciplinary field where information technology, computer science, social science, and sociology overlap. Role-Based Collaboration (RBC) has been proposed as a computational approach to facilitating collaboration. It utilizes roles as underlying mechanisms to support collaboration by taking advantage of roles. It is divided into several phases: role negotiation, role assignment, role execution, and role transfer. RBC and its related components are an abstract model, which is a perfect mapping for social activities, because Social and economic systems are typical collaboration systems. The Environments – Classes, Agents, Roles, Groups, and Objects (E-CARGO) model, which has been developed into a general model for complex systems have a good match for the requirements of computational social simulations. In this talk, we establish the fundamental requirements for social simulation and demonstrate that RBC, E-CARGO, and the subsequent Group Role Assignment (GRA) optimization model are highly qualified to meet these requirements. Based on E-CARGO and GRA, we present a new approach to social simulation with E-CARGO related components, models, and algorithms. This tutorial also illustrates several interesting case studies of computational social simulations.
|
|
Su-S2-WS2 Workshop Session, KEPLER |
Add to My Program |
Advances in the Development of Human-Machine Interaction (HMI) for
Connected and Automated Vehicles (CAVs) |
|
|
Chair: Li, Lingxi | Indiana University-Purdue University Indianapolis (IUPUI) |
|
16:30-18:30, Paper Su-S2-WS2.1 | Add to My Program |
Advances in the Development of Human-Machine Interaction (HMI) for Connected and Automated Vehicles (CAVs) |
|
Li, Lingxi | Indiana University-Purdue University Indianapolis (IUPUI) |
Good, David | Indiana University-Bloomington |
Na, Xiaoxiang | University of Cambridge |
Chen, Long | Institute of Automation,Chinese Academy of Sciences |
Cao, Dongpu | Tsinghua University |
Keywords: Human-Computer Interaction, Human-Machine Cooperation and Systems
Abstract: This workshop aims to host a panel discussion on recent advances in the development of Human-Machine Interaction (HMI) for Connected and Automated Vehicles (CAVs), with a particular focus on the practices, challenges, enablers and barriers in designing, developing, testing, and trialing all levels and aspects of HMI for CAVs. Our goal is to promote interdisciplinary research and international collaboration on the development and application of more intelligent, reliable, and secure HMI technologies for both on-road and off-road CAVs. The panel discussion will focus on but is not limited to the following core functionalities of HMI technologies for CAVs: navigation, driver alert, driver assistance and shared control, remote monitoring and taking-over control, passenger-vehicle communication, in-vehicle entertainment, user privacy protection, and road-user-vehicle interaction.
|
| |