KSEE 2024: Revisiting what does the progress in AI mean for Systems Engineering?

Title: Learning the language of nature: Generative AI for electron microscopy applications

Abstract: In this presentation, we will introduce Thermo Fisher Scientific and its electron microscopes that are being used in life science, material science and semiconductor applications. We will take a look at the various ways that Generative AI, beyond large language models, is being used to learn and understand ‘’the language of nature’’ in the context of for instance Meta’s Segment-Anything.  We'll delve into specific case studies using these technologies, showcasing how Thermo Fisher Scientific leverages AI in general and Generative AI in particular to bring a new perspective to applications and operation of electron microscopes. 

Short Bio: Dr. Remco Schoenmakers is Director Data and Artificial Intelligence in the R&D department of the electron microscopy division of Thermo Fisher Scientific. He holds a PhD in Astrophysics (1999) from the University of Groningen. He has worked at Philips Medical Systems on 3D CT and MRI visualization software before joining FEI in 2001 to work on image processing for 3D electron tomography software. He has developed the widely used Inspect3D electron tomography reconstruction package, and pioneered the EPU software package for Single Particle Acquisition, a technique that earned its academic developers the Nobel prize in Chemistry in 2017. He has fulfilled the roles of both group leader Application Software in Development and Digital Science in Research. He now is leading the strategy and development activities for Data Management and Artificial Intelligence infrastructure and applications in the development organization. 

Title: Accelerating Transformation in the Age of AI & Data

Abstract: This presentation delineates a part of the ongoing journey of digital transformation in Aker BP,emphasizing the need for a strategic, clear vision towards a modernized digital state. It highlights the primary challenges encountered in transforming supply chain operations, particularly focusing on the complications arising from unstructured data. This often results in manual labor, delivery delays, and an increase in master data duplication. To address these issues, an innovative solution is presented in the form of Cognite’s AI Document Parser. This AI-based tool streamlines data processing, thereby reducing waste, enhancing data quality, and ensuring consistent, error-free outputs necessary for effective digital transformation. The presentation underscores that digitaltransformation is not a finite project but a perpetual process that demands continuous commitment, adaptability to change, and a strategy flexible enough to meet evolving technological and business landscapes. Success in this journey is tied to maintaining a focus on the ultimate goals, adapting processes, and shifting organizational culture to fully leverage the potential of AI and data-driven technologies

Short Bio: Karen V. Czachorowski is an accomplished supply chain and digital transformation professional with a strong background in supply chain and the oil and gas industry. She is originally from Brazil and currently resides in Norway. Karen received her Bachelor of Business Administration degree from Paraná Social Studies Foundation/FESP institute, Curitiba, Brazil, in 2011. She went on to complete her Master of Science in Maritime Management from the University of South-Eastern Norway, Horten, Norway, in 2017. She later earned her industrial PhD degree in nautical operations/systems engineering from the same university in 2022. Karen has over a decade of experience in various supply chain roles across different industries, with a particular focus on the oil and gas sector. Since 2018, she has been leading supply chain improvement and digital transformation projects, including platform integrations, currently serving as a IT & Data Delivery Lead at Aker BP in Oslo, Norway, and is a public speaker in international events.

Title: Overcoming Obstacles in Developing Enterprise LLM Applications

Abstract: In this talk, Maged Helmy shares his journey of implementing Large Language Models (LLM) systems at the enterprise level. Over the past year, he has successfully implemented 8 LLM systems, adding to the dozens of AI systems he has integrated over the past 5 years. He discusses the unique challenges faced in integrating LLMs into existing business workflows and offers practical solutions to overcome these obstacles. He also highlights the advantages and limitations of LLM systems and provides advice on building your own. The talk covers a wide range of topics, including the challenges of application paradigm, model reliance, security/compliance, technical domain adaptation, cost operations, and more. It also explores various strategies for enhancing the quality of Retrieval Augmented Generation (RAG) architecture-based applications.

Short bio: Maged Helmy, who earned his PhD in AI from the University of Oslo (UiO), is an Associate Professor of Software Architecture in the Computer Engineering Department at the University of South-Eastern Norway (USN) in Kongsberg. Over the past 8 years, he has provided AI consultancy services to several organizations, including the Norwegian Digitalisation Agency (Digdir), the Norwegian Board of Health Supervision (Helsetilsynet), and Cognite.

Title: Can generative systems design solve the flaws of MBSE?

Title: Exploring large language models for legacy software

Abstract: Imagine a future where generative AI can provide intelligent assistants for software and system engineering, solving problems like knowledge management, root-case analysis, code understanding, and restructuring. TNO-ESI (NL) and Philips (NL) are charting this future together. This presentation will show our initial results of using large language models to boost state-of-the-art code analysis and restructuring methodology to tackle the ever-increasing complexity problem of dealing with legacy software.

Short bio: Rosilde Corvino is a Senior Research Fellow with experience in embedded software development and project management. She earned a Master of Science in Electronic Engineering from the Politecnico di Torino and a Ph.D. in micro and nanoelectronics from the Universite Joseph Fourier in Grenoble. 

Throughout her career, Rosilde has worked in processor design, compiler design, and software development. She has held positions at INPG and INRIA in France, as well as at TU/e, Intel, and TNO-ESI in the Netherlands. 

Currently, at TNO-ESI, Rosilde leads research activities to develop effective and efficient methods for managing software legacy, including exploring using large language models (LLMs) for knowledge management, code analysis, and restructuring.

Title: Collecting user needs using rapid prototyping - case study in the power production industry

Abstract: The master thesis presented was written in the power production industry in Norway, where the student has worked for more than ten years. In this presentation there is a short introduction to Sira-Kvina and a brief description of their digitalization process to gain improved access to data from their systems thus enable data driven decisions. The master thesis describes rapid prototyping to collect user needs for data access. The dashboard prototypes show live operational data from the power plants. To engage the stakeholders the student used rapid  rototyping as a technique. The student created prototypes of dashboards showing live operational data from the power plants. The data behind these conclusions comes from a literature study and interview with stakeholders in the company. The presentation finishes with a short description of Sira-Kvina machine learning initiative.

Short bio: Geir Aslak Kinden works at Sira-Kvina kraftselskap as a maintenance engineer. Kinden started working at Sira-Kvina kraftselskap in 2000 and received his apprenticeship as an industry mechanic in 2002. He then started working in the oil and gas industry onshore. He got a technical college degree in machinery from the Stavanger Offshore Technical College in 2005. After technical college he continued working in the oil and gas industry, but this time offshore. He rejoined Sira-Kvina in 2011. While working as an engineer at Sira-Kvina he took his second apprenticeship in 2020 as an energy operator. Kinden holds a bachelor’s degree in mechanical engineering from NTNU in 2016. He Started his MSc in Systems Engineering at the University of South-Eastern Norway (USN) in 2021.

Title: Navigating the New Frontier: Opportunities and Challenges of AI in Technology Development

Abstract: This presentation touches upon the dynamic landscape of technology development shaped by the rapid evolution of Generative Artificial Intelligence (AI). AI presents a myriad of opportunities for innovation, efficiency, and problem-solving across diverse sectors, but with these opportunities comes a whole set of new challenges.  The key focus areas include the enhancement Ai can give to development work, opportunities for the integration of AI in processes. Alongside the opportunities, the talk will also examine some of the challenges posed by AI, such as ethical considerations, GDPR, AI-ACT, and cybersecurity risks. The presentation aims to provide a balanced perspective on how to use the potential of AI while mitigating its risks.