The 12th PLAN-E plenary was hosted by CSC (Finland) and was open to all PLAN-E members.
Registration: Registration is closed!
Date: 24 November (13.00-16.45 CET) and 25 November (09.00-16.30 CET)
Organization committee: Leif Laaksonen, Kimmo Koski, Per Öster (CSC), Tom Bakker and Patrick Aerts (PLAN-E)
During the breaks the following breakout rooms were available for ad-hoc discussions and commenting: (1) EOSC, (2) Digital Health, (3) Artificial Intelligence, (4) Scientific Software Development, (5) Exascale & (6) Quantum Computing
Programme Tuesday 24 November 2020 (Day 1)
Moderator for Day 1: Sara Garavelli (CSC)
|13:00-13:10||Practical and agenda matters|
by Tom Bakker (secretary PLAN-E) – Presentation
|13:10-13:20||Welcome by the local host|
by Kimmo Koski (director CSC) – Presentation
|13:20-13:30||Introduction to the plenary|
by Joris van Eijnatten (chair PLAN-E)
New developments in HPC
|13:30-14:15||Towards the exascale: LUMI/EuroHPC status and future plans|
by Kimmo Koski (CSC, Finland) – Presentation
|14:45-15:30||Some peculiarities of HPC at CEA – From in-depth co-design to high-end infrastructure and services deployment for industry and research|
by Jean-Philippe Nominé (CEA, France) – Presentation
|15:30-16:15||Current developments and state of play – the perspective of the European Commission|
by Juan Pelegrin/Oscar Diez (DG CONNECT) – Presentation
|16:15-16:45||Plenary discussion on New Developments in HPC with a summary|
Moderated by Patrick Aerts (eScience Center, the Netherlands)
|End of Day one|
Programme Wednesday 25 November 2020 (Day 2)
Moderator for Day 2: Joris van Eijnatten (Netherlands eScience Center)
AI: Privacy, Trust & Transparency
|09:10-09:40||Digital health – handling of sensitive biomedical data|
by Megi Sharikadze/Dieter Kranzlmüller (Leibniz Supercomputing Centre, Germany) – Presentation
|09:40-10:10||Artificial intelligence and biology – Role of structured data in knowledge-building|
by Tommi Nyrönen (CSC, Finland) – Presentation
|10:30-11:00||AI and health (data) – Privacy, transparency and responsibility: an overview of the most crucial legal aspects|
by Elisabetta Biasin (KU Leuven Centre for IT & IP Law, Belgium) – Presentation
|11:00-11:30||Plenary Discussion on AI: Privacy, Trust & Transparency|
Moderated by Per Oster (CSC)
Presentations on – Research Software Engineering in Europe
|11:30-12:00||RSEs: collaboration between researchers and software experts,|
by Paul Richmond (University of Sheffield, UK) – Presentation
|13:00-13:30||The CodeRefinery project: teaching research software development and connecting Nordic research software engineers|
by Radovan Bast (UiT The Arctic university of Norway) – Presentation
|13:30-14:00||RSEs at Dutch research organizations|
by Joris van Eijnatten (Netherlands eScience Center) – Presentation
|14:00-14:30||Plenary discussion on Research Software Engineering in Europe|
Moderated by Joris van Eijnatten
The future of Quantum Software
by prof. Harry Buhrman, University of Amsterdam and CWI (national research institute for mathematics and computer science), and director of QuSoft (Dutch research center for quantum software), The Netherlands – Presentation
|15:45-16:15||Discussion on the future of PLAN-E|
|16:15-16:30||Closing words by host, next host (Hungary) and PLAN-E – Announcement János Mohácsi (KIFU, Hungary)|
Presenter bios & abstracts
Dr. Kimmo Koski (Managing Director, CSC – the Finnish IT Center for Science) started in his current position at CSC in August 2004 with a mission to support Finnish research through providing world-class e-infrastructure and related services. This has included investments in national resources and active participation in European initiatives. One example of the recent projects is the eco-efficient datacenter in a former paper-mill in Kajaani, which has been running in production since 2012. In 2020 CSC’s datacenter in Kajaani will host EuroHPC pre-exascale system provided by a 10-country LUMI consortium coordinated by Kimmo Koski.
Presentation abstract: The talk will discuss the future plans in EuroHPC and LUMI consortium on the way towards exascale – as also ways to develop European competiveness and collaboration through building an efficient ecosystem around the EuroHPC concept.
Dr. Jean-Philippe Nominé joined CEA HPC division in 1992, where he held different managing positions in HPC software development, before getting involved in European activities from 2007 onwards. Dr. Nominé was involved in PRACE European HPC infrastructure since the beginning of its preparation, and was a Member of PRACE aisbl Board of Directors in 2010-2011. He was then ETP4HPC Office manager between 2012 and 2019 and is now a member of ETP4HPC Steering Board, and of EuroHPC Research and Innovation Advisory Group (RIAG). At CEA he manages HPC strategic collaborations (EU and international). J.P. Nominé graduated from Ecole Polytechnique (engineer degree) and holds a PhD from Université Pierre-et-Marie-Curie (Paris, 1991). CEA is both an advanced user of HPC for its programmes and missions, and an HPC « provider » – a supercomputing centre operator and provider of related services, as well as an active HPC R&D player and HPC technologies developer.
Presentation abstract: We will present the different facets of HPC at CEA, and some of their interactions with the European HPC ecosystem:
– Setup and operations of a supercomputing ‘complex’ comprised of 3 large machines and two facilities on the same site near Paris, for research, industry and defence uses at the French and/or European levels
– Development of HPC hardware and software technologies, in close relationship with industrial suppliers, with in particular a long standing « deep » co-design approach of supercomputers
– Development and use of large scale numerical simulation, data analytics and machine learning applications, in close relationship with many scientific communities and industrial partners
Head of Sector of Exascale Computing in the HPC & Quantum Unit.
Recent developments on HPC in Europe and the plans for the future.
AI: Privacy, Trust & Transparency
Megi Sharikadze (PhD in Human and Animal Physiology, 2004), joined LRZ in 2014 as an academic project manager. Since then she has been acting in her different functions dealing with academic project acquisition and execution as well as science management and coordination. Drawing upon 15 years of experience with European research funding she as an EU Research funding advisor largely contributes to LRZ’s European proposals and projects.
In her previous occupation at University Hospital Regensburg, acquisition and management of European health-related projects was her main responsibility (till 2014). E.g., EC-FP7 large-scale collaborative project LipidomicNet (GA: 202272) was supported and managed by her (2009-2014). Prior to these activities, she conducted neuroscientific research 1998-2009 (at the end as a senior researcher scientist at I. Beritahsvili Institute of Physiology (Tbilisi, Georgia) and guest scientist of university of Federal Armed Forces (Munich, Germany).
Nowadays, in parallel to consulting researchers and partners in research fundraising, she pays special attention to Digital Health domain. With this respect she consolidates efforts of LRZ’s BioMedICT task-force, and especially supports community building and networking activities at local, national as well as on European levels (also via active participation in Plan-E, e-IRG and other initiatives).
Presentation abstract: With wide acknowledgment of essential footprints of data science, AI and computerized reasoning in modern biomedicine, data collection and sharing as well as next generation analytic methods, applications and respective digital infrastructures are rapidly advancing. These components attract increased interests not only of biomedical professionals and ICT experts, but also of general public, politics and society. In context of ongoing COVID-19 pandemic, interests in Digital health issues grow and stay overwhelming.
In the “AI: Privacy, Trust & Transparency session” of the 12th Plenary meeting of Plan-E, under umbrella of Digital health topic, several aspects of biomedical data handling will be addressed, drawn upon experiences in current initiatives and projects at LRZ. Among the lessons learned, one should emphasize is that successful data protection is not based on pure technical issues, but also requires medical practitioners to adapt their way of using IT infrastructures.
Tommi Henrik Nyrönen PhD AdjProf is the Head of the ELIXIR Finland Node operated by CSC – IT Center for Science in Finland. His role is enabling research, focusing on services for big life sciences data in ELIXIR, European Life Science Infrastructure for Biological Information (http://www.elixir-europe.org). In 2019, EU member states and Commission elected him to co-lead the technical interoperability of the EU 1 million genomes initiative. He also leads the Data Use and Researcher Identities workstream in the Global Alliance for Genomics & Health GA4GH, and a co-lead of the European ELIXIR Compute Platform, adjunct professor in computational drug design at the University of Helsinki, and a past by-fellow in bioinformatics at Churchill College, University of Cambridge.
Presentation abstract: Even in the age of emergence of artificial intelligence, the development of new individual prediction algorithms follows the old proverb: poor data, poor models. Learning does not happen without access to proper learning materials. The same principle applies to both humans and machines. Machine learning requires different, machine-readable, representation of high-quality data. Therefore, the attention of researchers is increasingly turned into solving the problem of how to gain access to high-quality datasets.
In this talk, I will introduce European data resources and data access processes in biology. In particular I will discuss how to facilitate access to large-scale sensitive genome and image data consented for research, as well as processes to gather and broker computational access to the data using international standards. The goal is to discuss the importance of sustained high-quality data sources to develop new knowledge, algorithms and models to understand human health.
Elisabetta Biasin is a Researcher in Law at the KU Leuven’s Centre for IT and IP Law (CiTiP) where she researches in the field of eHealth, privacy and data protection, cybersecurity, and medical devices law. Elisabetta is currently involved in EU studies and several EU funded projects, including Made4You, SAFECARE, PharmaLedger – where she investigates perspective concerning healthcare data sharing, medical devices privacy and cybersecurity. Before joining CiTiP, Elisabetta worked as a Legal Advisor on Privacy and Data Protection for Deloitte Legal Italy. Before, she was a Trainee for European Digital Rights (EDRi) in Brussels and the Bolzano Criminal Court in Italy.
Presentation abstract: The new advent of AI technologies in healthcare lead healthcare stakeholders – including researchers – to investigate their many related legal facets. Novel applications of AI in healthcare are wide-ranging from data-driven analytic tools for personalised medicine, to assistive technologies, robotics surgery, to name a few. The use of such tools, however, requires a thorough evaluation of privacy and data protection, cybersecurity and medical devices legal aspects in their intertwined aspects with transparency and responsibility. This presentation will guide through the most relevant legal elements in the EU, as understanding and adhering to requirements thereof are essential not only to guarantee the ensure patient’s privacy and medical confidentiality but also to promote trust in healthcare and research institutions.
Research Software Engineering in Europe
Dr Richmond is a Research Software Engineer (RSE) and director of the Research Software ENgineergin Group and the University of SHeffield, UK. He is currently supported by an EPSRC Early Career, Research Software Engineering (RSE) Fellowship and is actuating President of the Society for Research Software Engineering. He leads the Research Software Engineering group at the university of Sheffield which more broadly embeds software engineering skills and software engineering best practice into interdisciplinary collaborations to facilitate better and more sustainable research. As part of his fellowship, Dr Richmond facilitates the use of accelerated architectures such as Graphics Processing Units (GPUs) to accelerate scientific discovery. He is developing software techniques, a provision of skills and training material to help drive the use of accelerators into mainstream science and engineering. More broadly, Dr Richmond advocates for the role of Research Software Engineers and is developing a sustainable pathways to supporting a wide range of research software through his RSE group and within the wider RSE community.
Presentation abstract: The Research Software Engineering movement is at the forefront of change in research practice ensuring modern computational research is reproducible. They key to empowering Research Software Engineers lies within the community itself. This talk will outline the growth of this community and recognition of the RSE role in the UK and beyond. It will give insight into providing career structures within academic/research institutions and describe some of the successes of the RSE Society as well as some future plans for developing the RSE role.
Radovan Bast is a research software engineer with background in theoretical chemistry. He’s worked in France, Stockholm, and now Tromsø at the border between science, software, and computational support and enjoys supporting multi-disciplinary research. He now works as part of the Sigma2 metacenter at the University of Tromsø, Norway, and leads the CodeRefinery project
Presentation abstract: In my presentation I will discuss how we grew the CodeRefinery project over the past 4 years and taught hundreds of students and researchers across all disciplines in best practices in reproducible research software engineering. I will also discuss our efforts to build a community of research software engineers in the Nordics.
Joris van Eijnatten
Joris van Eijnatten (PhD 1993) is general director of the Netherlands eScience Center and professor of digital history at Utrecht University, the Netherlands. His numerous publications include books and articles on toleration, Dutch religious history, the poet Willem Bilderdijk, sermons, the justification of war, media history and conceptual history. A digital historian, Van Eijnatten specialises in applying digital methodologies to newspapers, periodicals, parliamentary records and other historical data, with a focus on tracing changing concepts over time.
Presentation abstract: At the present moment probably only a handful of people at any Dutch research organization would recognize the term ‘RSE’, even if the abbreviation were spelt out in full. The underlying problem, however, is not so much that the term RSE in not common usage yet, but the general lack of recognition of the people who identify with what the term RSE stands for. In this respect the situation in the Netherlands is not very different from anywhere else. One of the ambitions of the Netherlands eScience Center is to push forward recognition of RSEs on the local and national levels, and in this way contribute to change on the European level. In this talk I will briefly discuss the current situation in the Netherlands, moving outwards from the Netherlands eScience Center itself towards more general developments in research, research funding and research policy.
Keynote on Quantum software: Harry Buhrman
Harry Buhrman is professor of algorithms, complexity theory, and quantum computing at the University of Amsterdam (UvA), group leader of the Algorithms & Complexity group at the Center for Mathematics and Informatics (CWI), and executive director of QuSoft, a research center for quantum software, which he co-founded in 2015. In 2020 he was elected as a member of the Royal Netherlands Academy of Arts and Sciences (KNAW). In the mid 1990’s he built the quantum computing group at CWI, which was one of the first groups worldwide and the first in the Netherlands working on quantum information processing. Buhrman’s research focuses on quantum computing, algorithms, and complexity theory. He co-developed the area of quantum communication complexity (distributed quantum computing), and demonstrated for the first time that certain communication tasks can be solved (exponentially) more efficient with quantum resources. This showed that quantum computers can not only speed up computations, but also make certain communication task more efficient. The latter opened up a whole new research and application area of quantum information processing. Buhrman co-developed a general method to establish the limitations of quantum computers, and a framework for the study of quantum query algorithms, which is now textbook material. Buhrman is coordinating and lead principle investigator on the Quantum Software Consortium (QSC).
In the past years Buhrman established two no-go Theorems for quantum cryptography. He showed that secure multi-party quantum-computation and position-based quantum-verification are unconditionally impossible. In order to construct position-based quantum-protocols that are secure under reasonable assumptions on the attacker’s resources, he developed a new model of communication complexity called the garden-hose model. He furthermore employed ideas from quantum information to solve a 35-year-old problem in functional analysis. Recently he established a general method to construct a Bell inequality with a quantum violation from a quantum communication complexity advantage.
Presentation abstract: The talk will start with a short introduction to quantum computing with a focus on applications and quantum software. We will then discuss the state of the art of the field and what the challenges are that we have to face in the coming years regarding the interactions between science, society and industry.