[ECSS 2024] Presentations of the Keynotes Session (29 October 2024)

Session Chairs: •Jean-Marc Jézéquel, Informatics Europe / IRISA / University of Rennes •Enrico Nardelli, University of Rome "Tor Vergata" (Italy) 20th European Informatics Leaders Summit (ECSS 2024) Keynotes Session "Reflecting on the past and the future of Informatics in Europe“ 29 October 2024 Reflecting on the past and the future of Informatics in Europe –a personal view Anne-Marie Sassen –HoU, European Innovation Council and SME Executive Agency 22 My personal history of informatics in Europe My job Trend in the European Commission 1997 - 2005: Involvement in European projects related to software engineering, while working for ATOS Spain Software engineering important as a topic on its own 2005 –2015: Project officer in Software and Cyberphysical systems units of DG INFSO/DG CONNECT. Software focus shifted towards cloud computing. Not pure research any longer but more policy actions. Software was “the missing KET” (Key Enabling Technology) according to some. 2016 –2022: Policy officer related to digitalisation of European industry Digitalisation of European Industry (Industry 4.0), European Digital Innovation Hubs, later AI for Europe, AI Act, AI Office 2022 –now: European Innovation Council Focus on deep tech, breakthrough innovation, bringing research to the market 33 The future policy context: Some words from Mario Draghi “First – and most importantly – Europe must profoundly refocus its collective efforts on closing the innovation gap with the US and China, especially in advanced technologies.” “Public sector support for R&I is inefficient due to a lack of focus on disruptive innovation and fragmented financing, limiting the EU’s potential to reach scale in high-risk breakthrough technologies.” “Universities and other research institutions are central actors in earlystage innovation, generating breakthrough research and producing new skills profiles for the workforce.” The future (2) •A strong focus on strategic technologies: ▪Advanced semiconductors technologies ▪Artificial intelligence technologies ▪Quantum technologies ▪Advanced connectivity, navigation, and digital technologies ▪Robotics and autonomous systems ▪Biotechnology ▪Advanced materials ▪Cleantech ▪… EIC Work Programme 2025 General Info day: 5 November Accelerator Challenges Info day: 6 November © European Union, 2021 Reuse of this document is allowed, provided appropriate credit is given and any changes are indicated (Creative Commons Attribution 4.0 International license). For any use or reproduction of elements that are not owned by the EU, permission may need to be sought directly from the respective right holders. All images © European Union, unless otherwise stated. Image sources: ©Tom Merton/Caia Image, #315243588; ©REDPIXEL, #220695664; ©Halfpoint, #180578699; ©bnenin #213968072; ©MyMicrostock/Stocksy, #3094437622021. Source: Stock.Adobe.com. Icons © Flaticon – all rights reserved. Thank you [email protected] Assured LLM-Based Software Engineering ECSS 2024 keynote Mark Harman 29th October 2024 Joint work with Nadia Alshahwan, Andrea Aquino, Jubin Chheda, Anastasia Finegenova, Inna Harper, Mitya Lyubarskiy, Neil Maiden, Alexander Mols, Shubho Sengupta, Alexandru Marginean, and Eddy Wang. A brief note about how I got here Challenges in Generating unit and integration tests Software Testing 101 Traditional testing pyramid Traditional testing pyramid Traditional testing pyramid Traditional testing pyramid Unit and integration distinction is blurry Traditional testing pyramid High compute cost Low compute cost Traditional testing pyramid High human effort Low human effort Traditional testing pyramid - in theory Few Many Traditional testing pyramid - in practice Some tests. Non trivial human effort Some tests. Sometimes unexpectedly high human effort Exploring e2e in more detail 5 Single user Multiple independent users Production Exploring e2e in more detail 5 Single user Multiple independent users Interacting user community Production Exploring e2e in more detail: New social testing 5 Single user Multiple independent users Interacting user community Production Exploring e2e in more detail: New social testing 5 Single user Multiple independent users Interacting user community Production Traditional e2e Stress Testing Beta testing Social testing Simulation Based Testing at Meta timeline Alpha Release Change request published Change in main branch Beta Release Find issues early in development process testing diffs and master builds Identify high impact issues by simulating production and prevent their release Automated Test Generation: next … unit tests! Forthcoming papers … three come along at once Published papers Assured LLM-Based Software Engineering arXiv and ICSE 2024 InteNSE workshop keynote Automated Unit Test Improvement using Large Language Models at Meta arXiv and submitted to FSE 2024 Industry track Observation-based unit test generation at Meta arXiv and submitted to FSE 2024 Industry track Published papers Assured LLM-Based Software Engineering arXiv and ICSE 2024 InteNSE workshop keynote Automated Unit Test Improvement using Large Language Models at Meta arXiv and submitted to FSE 2024 Industry track Observation-based unit test generation at Meta arXiv and submitted to FSE 2024 Industry track Automated Unit Test Improvement using Large Language Models at Meta Automated Unit Test Improvement using Large Language Models at Meta Automated Unit Test Improvement using Large Language Models at Meta Automated Unit Test Improvement using Large Language Models at Meta Automated Unit Test Improvement using Large Language Models at Meta Automated Unit Test Improvement using Large Language Models at Meta: results Conclusions Assured LLMSE is the new Genetic Improvement LLMs are the new GP So much exciting research to do Next steps Use high quality regression tests as the filter for LLMSE Details in Assured LLM-Based Software Engineering paper Teaching Dorothea Wagner Karlsruhe Institute for Technology Politics Teaching Politics Laurentius de Voltolina, 14th c. Teaching - My background: The German Council for. Science and Humanities (“Wissenschaftsrat”) - When Science advises politics (about informatics) - Dealing with mutual misunderstandings Topics: German Council for Science and Humanities •Founded in 1957 •Scientists (also some from industry), media, public figures •Federal minister of education and research and the ministers of the 16 federal states (“Länder”) •Advising politics on all aspects of the German science system • Evaluations, accreditations, university medicine,… German Excellence Initiative,... •Recommendations on current issues and challenges 2) Sovereignty and security of science in the digital space •Governance structures supporting digital empowerment •Professionalization and attractiveness of the employment sector •Raising awareness of dependencies and risks •Overarching cooperation, advisory services, competence centers,… •Permanent task that needs permanent financing →Recommendations with financial consequences are challenging! Source: ChatGPT Persistent decrease of student competences •Teacher training core task at the training institutions •Studies program integrating theoretical and practical phases •Single-subject teacher training for mathematics specialists •New ways for career changers •Shorten the training path for teachers •Eliminate gap between the two training phases •Increase attractiveness of the teaching profession 3) Teacher training in mathematics Source: ChatGPT 4) Perspectives of informatics •Pooling of forces, profile building at the various locations •Balanced funding policy •Measures in the competition for top talents • Attract more students, increase students’ success rate •Larger and broader range of study programs •Education in computer science at schools •More active role in communication with society, media and politics Source: ChatGPT Understanding and misunderstanding “Informatics”: Digitalization Source: Google Books Occurrence of the term per year Search results for “digital” on Google in 2006 –rectangular formats dominate Search results for “digital” on Google in 2016 –shades of blue dominate →unlimited sky,… 10 Years Classical view of informatics View and focus of politics Understanding and misunderstanding “Informatics”: Digitalization Understanding and misunderstanding Artificial Intelligence Understanding and misunderstanding Algorithms Source: The Telegraph Source: The Telegraph Understanding and misunderstanding Algorithms We strongly urge a clear distinction between “algorithms” and “artificial intelligence”, partly because it reflects reality, but mostly because the confusion of these two very different concepts in popular discourse muddies the already difficult discussion of the implications of “Digitalisierung”. For clarity, “algorithms” are a foundational concept in computer science, and refer to… In December 2019, the Department of Computer Science at ETH Zurich felt compelled to criticize the increasing equation of algorithms and artificial intelligence even within ETH. The reason for this was the draft of ETH Zurich’s new strategy plan. A letter to ETH’s administration states: AI and algorithms Algorithms for route planning in transportation (one of my research topics) Showpiece for AI?