Increasing numbers of objects in our day life are controlled by computers: phones, aircraft, cars, buildings, manufacturing machines, musical instruments, etc. In these so-called cyber-physical systems (CPSs), computers interact directly with the physical world through sensors and actuators. Those systems are becoming the key infrastructure and backbone of our society and are at the heart of revolutionary changes in our daily lives and economy.
The sophistication and complexity of CPSs keeps increasing, since they must realize more functions with limited resources, which makes them increasingly difficult to build and manage. In particular, the cyber (software) part of these systems is growing rapidly, and has become a key part in CPS, as they are the basis of operation for these systems.
Artificial intelligence (AI) has a fundamental influence on all areas of economy, administration and society. AI is now also affecting software engineering; providing robust approaches for software development in order to analyze and evaluate complex software and its development processes. Repository mining, machine Learning, big data analytics and software visualization enable targeted insights and powerful predictions for software quality, software development and software project management.
This workshop aims at gathering researchers and practitioners addressing the challenges induced by the software aspects of AI, in order to identify synergy, common problems, solutions and visions for the future of this area. Strong interactions among participants will be favored to provide constructive feedback for accepted workshop papers and develop future collaborations and community building.
Manoj Chinnakotla, Microsoft
Dr. Manoj Chinnakotla is a Senior Applied Scientist in the Artificial Intelligence and Research (AI&R) group at Microsoft, Bellevue, USA. He is also an Adjunct Faculty at the Language Technology Research Centre (LTRC), IIIT Hyderabad. He holds a Ph.D. in Information Retrieval (IR) and Natural Language Processing (NLP) from IIT Bombay where he was an Infosys Research Fellow. He has published in several reputed conferences and journals in AI, IR, and NLP such as SIGIR, ACL, AAAI, and IJCAI which have been cited more than 500 times. He was the recipient of the "Best Paper Award" at PAKDD 2017 conference for his pioneering work on detecting inappropriate content in search queries. During his current stint at Microsoft, his research has powered a lot of features across popular Microsoft products such as Windows, Cortana, Bing etc. which are being used by millions of people across the world. His current research interests include Question Answering, Entity Mining, and Web Ranking.
The workshop consists of a series of invited presentations. The presentations have been divided into tracks based on the application domains.
AI IN BIOENGINEERING
Bio-CPS are considered to be an integration of computational elements within biological systems. In a sense, Bio-CPS can be compared to the Cyber-Physical systems, in which the challenge is to make biological systems working along with computer systems. In biological systems the complexity is required in order to make the system more stable. By contrast, in cyber systems engineering, the complexity needs to be avoided as far as it is possible because it can bring instabilities. However, cyber systems complexity, even not desired, stems from the numerous interactions between components. Consequently, the main concern in Bio-CPS design is how to make these two kinds of systems coupling together to perform a common task with a high level of confidence.
BioCPS have far-reaching applications in real-life. Some of them include revolutionizing the healthcare system, medical instrumentation, medical technologies, and modelling diseases, infections, and the human immuno-response.
Veeky Baths, BITS Pilani, Goa Campus
AI IN CYBER-PHYSICAL SYSTEMS
While machine learning is widely practiced in cognitive domains, such as NLP, computer vision, speech recognition, etc, its impact on cyber-physical systems is only recently being understood. An important hurdle towards the deployment of machine learning in industrial systems is that machine learning has developed in Computer Science, whereas its applications in industrial domains require significant domain knowledge in disciplines that are remote from Computer Science. On the other hand, a large arsenal of ready-to-use machine learning tools and libraries exist (including ones supported by standard industrial modeling platforms like MATLAB), and therefore the domain expert can be readily trained to use these tools and techniques for advancing their industrial processes and design (R&D) activities.
Dominique Blouin, Telecom Paris
"Model Management for Architecture-Centric Development of Cyber-Physical Systems with Multi-Paradigm Modeling"
Multi-Paradigm Modeling (MPM) is an approach to tackle the complexity of Cyber-Physical Systems (CPS) by modeling every aspects of a system explicitly using the most appropriate formalism(s) at the most appropriate level(s) of abstraction. With this approach, several modeling paradigms and their supporting formalism must be jointly employed to cover the heterogeneity of domains and different levels of abstraction of CPSs. Managing these models is therefore essential to ensure that their interplay and the activities performed on them do not lead to inconsistencies, which can be the source of costly errors introduced during concurrent engineering.
Despite its importance, model management is not yet well addressed. In this presentation I will introduce multi-paradigm modeling, our current effort on providing a formal definition for it and our research on model management applied to the Architecture-Centric Virtual Integration development Process (ACVIP) centered on the SAE AADL architecture description language. I will focus on the challenges faced when applying model management techniques to industrial modeling settings and present our ongoing work to address these problems. I will also briefly discuss some perspectives on bringing the assets of AI to MPM.
AI IN EDUCATION TECHNOLOGY
Education has been accepted as a project for social and political transformation, with the development of each individual not only for her economic gains, but also for building a just and humane society. It also needs to promote awareness and build agency for sustainable development and harmonious co-existence. Global policy documents such as the Education for All, Millennium Development Goals and the Sustainable Development Goals, emphasize universal education.
Poor investment in education results in poor quality of teacher education, and inadequate academic infrastructure. Teachers are unable and/or unwilling to provide support to the learning processes. In this context, digital technologies (aka Information and Communication Technologies, or ICT in short), are sometimes seen as a solution that can address curricular resource shortage, teacher shortage and teacher quality.
Applying both AI and CPS concepts in developing educational technologies can significantly improve these implementations and ensure efficiency. This is an up and coming field which includes the development of hardware and software facilitating both teaching and learning on various scales. Radical new ideas and platforms are coming up every day and modelling the interactions between the human component and the system can be effectively modeled using AI and CPS concepts.
Mattias Ulbrich, Karlsruhe Institute of Technology
To make the workshop wholesome and appropriate for beginners in Machine Learning and Artificial Intelligence, we have a series of lectures over the course of the day, to facilitate learning.
FOUNDATION OF MACHINE LEARNING :
"Beyond Hyperparameter tuning: Insights of Deep Learning Methods in AI"
Snehanshu Saha holds Masters Degree in Mathematical and Computational Sciences at Clemson University and Ph.D. from the Department of Applied Mathematics at the University of Texas at Arlington in 2008. He was the recipient of the presitigious Dean's Fellowship during PhD. After working briefly at his Alma matter, Snehanshu moved to the University of Texas El Paso as a regular full time faculty in the Department of Mathematical Sciences. He is an Associate Professor of Computer Science and Anuradha and Prashanth Palakurthi Centre for Artificial Intelligence Research (APPCAIR) since 2019 and heads the Center for AstroInformatics Modeling and Simulation (CAMS). Dr. Saha is also the Director-AI Research at HappyMonk Technologies, Bangalore.
The workshop is scheduled for 25 February 2021.