2017 Annual Report
Summary 2017 has been a highly active period for Amsterdam Data Science (ADS) with many key goals achieved. Our network organisation has continued to grow with new knowledge institutions joining, collaborations with industry established, and a broadened international network. This report summarises our continued growth and our heightened involvement within the Amsterdam region (our 2016 report can be viewed here: https://tinyurl.com/y77y6bhe). The report also updates on highlights of the Amsterdam Academic Alliance Data Science (AAA-‐DS) programme, as we become one joint initiative. 2017 highlights (in brief) Research
• New ADS Partners: Faculty of Sports and Nutrition (AUAS); Department of Mathematics, Faculty of Economics and Business, Amsterdam Center for Business Analytics (VU)
• New ADS Labs and company collaborations with Bosch (Germany), the Police (NL), and VIVAT (NL)
• Collaboration agreements signed with international knowledge institutes: Insight (Ireland’s data science research centre) and Tsinghua University (China)
• Awarded five joint international exchanges for PhD and postdocs with our collaborators in Ireland, Tsinghua and Wuhan
• Joint professor appointment in Urban Analytics between ADS partners AUAS and CWI Education
• Launch event of the Amsterdam School of Data Science (March 2017, 400+ attendees) • Launch of the Amsterdam Data Science Minor (September 2017, course over-‐subscribed) • Double increase in student numbers for the Data Science Master Track • Received the highest public vote at the Computable Awards for ICT Educator of the Year
Dissemination • Hosted 34 Meetups for students, researchers and industry experts on a range of data
science topics (23 Amsterdam Data Science, 11 from Search Engines Amsterdam) with an average of 100 up to 250 visitors each meetup. Including our end of year highlights event where we had 15 x four minute presentations from our partners
• Series of articles in ‘Het Financieele Dagblad’ on ADS collaborations with Elsevier, Tata Steel, Qualcomm, Sandd, Philips and Databricks
• Our Lab with the National Police lab featured heavily in national media outlets • Significant activity on online communication channels: Newsletter (+1.000); LinkedIn (+300);
Twitter (+1.000); Facebook (+100) Who we are ADS is a network organisation that coordinates shared research across knowledge institutes and industry/academic boundaries, initiated in 2013 by the Amsterdam-‐based knowledge institutes: Amsterdam University of Applied Sciences (AUAS), Centrum Wiskunde & Informatica (CWI), University of Amsterdam (UvA), and Vrije Universiteit Amsterdam (VU). ADS brings together over 600 scientific staff, at all levels, for example:
• Master students who spend a day a week on a specific ADS research project (9) (see project summaries here: https://tinyurl.com/y9jqc7pt);
• PhD and Postdoc researchers on specific AAA-‐DS projects (16) (see Table: https://tinyurl.com/y8meekd4);
• Associate and full Professor level researchers (over 100) from multiple research disciplines. Mission Amsterdam Data Science accelerates data science research, education and valorisation by connecting, sharing and showcasing world-‐class technology, expertise and talent from Amsterdam on a regional, national and international level. Our activities enable business, society and science to better gather, store, analyse and present data in order to gain valuable insights and make informed decisions.
Finance In 2017, each official ADS partner: Faculty of Digital Media and Creative Industries (AUAS); CWI; Informatics Institute (UvA); Amsterdam Business School and Amsterdam School of Economics, Faculty of Economics and Business (UvA); Department of Computer Science (VU); Department of Mathematics, Faculty of Economics and Business, Amsterdam Center for Business Analytics (VU), contributed 10K in cash and 40K in kind. The 10K cash contributions are utilised to support project staff (project management and business relations), hosting of regular seminars and events, and marketing and website maintenance costs. This is both for Amsterdam Data Science and the Amsterdam School of Data Science (our education platform). Within ADS, there is also the AAA funding stream, this is a joint data science initiative of UvA and VU, which aims to intensify collaboration between the two. Additionally it is also a collaboration involving AUAS, CWI, and VUmc, and industry, namely ORTEC and Philips. The AAA-‐DS projects are unique as they span multiple research domains (business, humanities, informatics, life sciences and social sciences), faculties and universities. There are 12 postdocs and 4 PhD students, each on a specific data science project within this 3-‐year programme. Table 1 summarises each AAA project with highlights and future project plans. In brief:
• 3.069K was the original funding amount awarded by the Boards of UvA and VU. In addition to this original amount the following has been secured:
• 1.255K for AAA projects with ABS, AUAS, CWI, Philips, ORTEC and VUmc; • 4.272K in NWO grants to AAA researchers for project collaborations with Ahold Delhaize,
Beeld en Geluid, Elsevier, and KNAW; • 8.596K in total cash injection to the AAA programme.
2017 highlights (in detail) Establishing new partners New knowledge institutes join ADS: In January 2017, the Department of Mathematics, Faculty of Economics and Business, Amsterdam Center for Business Analytics (ACBA) of the VU joined ADS. This partnership has already led to new projects being set-‐up, specifically in the domains of HR and big day analytics. At the end of 2017, the Faculty of Sports and Nutrition of the AUAS confirmed their intent to become an official ADS partner as of January 2018, and we are also in final talks with the Faculty of Social and Behavioural Science (UvA) who are keen to become an official partner shortly. New ADS labs: In 2017 alone, three new ADS Labs launched which will host over 30 joint PhD and postdoc positions between the University and the respective company:
• DELTA: A collaboration with Bosch on deep learning in the automotive industry and computer vision amounting to 10 positions
• The Police: A collaborative project which will use computer technologies to gain information from seized mobiles and computers 3 positions and 100K of yearly funding
• VIVAT: A collaboration on advanced analytics and Insurtech to 3 positions, amounting to 1,2 million Euro funding
New international collaborations: In 2017, ADS set-‐up new collaborations with Insight (data science research centre Ireland) and Tsinghua University (China). These collaborations will focus on joint research projects, education, and student and staff exchanges. In February, Insight researchers visited Amsterdam to discuss concrete collaboration plans and sign a memorandum of understanding. The one-‐day workshop involved over 20 researchers where each presented areas they would like to collaborate on. In June, four ADS professors visited Tsinghua to discuss collaboration plans. In October, there was a return visit from six Tsinghua professors to Amsterdam where further collaboration plans were discussed and an agreement signed. ADS awarded five exchange visits (commencing early 2018) with Insight, Tsinghua and Wuhan for PhD and Postdoc researchers. These exchange grants are jointly funded by the AAA programme and our international collaborators. A call for the second round of exchange visits has recently been announced, the visits will take place in the second half of 2018. In recent months a number of meetings with Beijing University of Technology and Columbia University have been held and it is planned for these collaborations to be formalised shortly. Strengthening the Amsterdam knowledge institutions collaborations AAA-‐DS: UvA hosts 8 positions and VU also 8, with the individual positions bridging universities, thus maximising collaboration opportunities. Eight positions span research domains (e.g., informatics/social sciences, business/informatics) and 2 involve local industry partners (ORTEC, Philips). AAA-‐DS also partners with other key Amsterdam academic partners: AUAS, CWI and VUmc. Research assistants: ADS has enabled 9 research assistant projects, where Master students spend 1 day a week on a specific project, with each individual project spanning either two knowledge institutes (AUAS, CWI, UvA, VU), or partnering with local industry (Dagblad De Correspondent), or a startup (CTcue). These represent new collaborations that would not exist without the ADS funding. Linking with related data science initiatives Responsible Data Science (RDS): Lead ADS researchers Maarten de Rijke and Frank van Harmelen coordinate the RDS team, a collaboration of 11 knowledge institutions across the Netherlands. Under the RDS theme, at Amsterdam we have coordinated six Meetup events this year, with speakers attending from across the Netherlands to present at this event. This has enabled this important topic to receive greater attention as evidenced through the VSNU call, which ADS researchers are closely involved in. Amsterdam Data Science also plays an active role in a number of national data science and computer science initiatives including the Big Data Alliance, Data Science Platform Netherlands, and the ICT Research Platform Netherlands. Promoting data science education in Amsterdam Amsterdam School of Data Science Events: In March we held the launch event of the School, an umbrella term to encompass the data science education offering from the Amsterdam knowledge institutes. The event, opened by the deputy mayor of Amsterdam, evidenced the high interest by the turnout of over 400 attendees. The School website is a platform for all course and programme options (more than 260) in data science and at all levels: Bachelor, Master, Postgraduate and Professional/Executive education. In this initiative, ADS have intensively marketed Amsterdam’s strong position in terms of data science study to the external community. In November, ADS, through Marc Salomon, hosted a half-‐day conference with Coursera (San Francisco) on digital transformation, the future of work and lifelong learning. There were a number of keynotes, for example from the CEO of Coursera, and panel discussions on ‘What are the skills of tomorrow needed in data?’ The event was attended by over 150 people from a broad range of
backgrounds: current data science students, those looking to begin their data science study, those looking to re-‐train, and representatives from the city and local industry. Master -‐ Data Science Track: ADS researchers have been actively involved in teaching this successful track and its development for the second year. Enrollments for year 2 have increased significantly from 35 (year 1) to 75. Minor -‐ Amsterdam Data Science: ADS researchers have set-‐up a new minor (started September 2017) to meet the high demands of those wishing to study data science. The minor has received significant interest since its launch and was over-‐subscribed very quickly. The minor will equip students with the skills to enroll in the data science master track. Both the minor and track are joint collaborative efforts of the Amsterdam knowledge institutions. Education awards: Amsterdam School of Data Science was nominated for the Computable Awards ICT Educator of the year (October 2017) and received the highest public vote. Disseminating our research to the outside world Events: ADS have two active local Meetup communities with over 5,500 members. ADS organises seminars and events (2 or 3 each month), which is an opportunity for our researchers to showcase their projects to key stakeholders. In 2017 we hosted over 30 events (see summary list below), with some events attracting audiences of up to 500 attendees. We also invite external speakers to present and have had speakers from Universities across Europe, China and the U.S., and speakers from industry including Amazon and Microsoft. In June, we hosted an event for 20 Amsterdam-‐based data science startups, with the aim to pitch in particular to our community. Communication: A monthly newsletter is mailed to our internal and external community (1000 subscribers), to highlight our research further and enhance its reach. We include research highlights, upcoming events and calls in the newsletter. In addition as further communication channels we have two websites (one for ADS and one for the School), two Twitter accounts (1300 followers), a LinkedIn company page and a Facebook page. Press: A series of seven weekly articles in ‘Het Financieele Dagblad’ on ADS collaborations with Elsevier, Tata Steel, Qualcomm, Sandd, Philips, Databricks, plus a summary article, were released at the start of 2017, which attracted great attention for the ADS research. The Police Lab has attracted a lot of attention and even made it to the national news of NOS and other media channels. ADS Management Team and Supervisory Board Members in 2017 Management Team: Nanda Piersma, Hajo Reijers, Maarten de Rijke (Chair), Marc Salomon. Supervisory Board: Henri Bal (Copmuter Science, VU), Peter Boswijk (Business, UvA), Lynda Hardman (CWI), Ger Koole (Business, VU), Sabine Niederer (AUAS), Max Welling (Informatics, UvA). Our Communication tools: Amsterdam Data Science Website: http://amsterdamdatascience.nl Email: [email protected] Meetup for events: https://www.meetup.com/Amsterdam-‐Data-‐Science/ Meetup (2) for events: https://www.meetup.com/SEA-‐Search-‐Engines-‐Amsterdam/ Twitter: https://twitter.com/ams_ds LinkedIn: https://www.linkedin.com/company-‐beta/10548382 Amsterdam School of Data Science Website: https://www.schoolofdatascience.amsterdam Email: [email protected] Twitter: https://twitter.com/AMS_DataSci Facebook: https://www.facebook.com/amsterdamschoolofdatascience/
For 2018 and beyond ADS strategic directions 2018-‐2022: Our 5 year strategic directions encompass Research, Education and Entrepreneurship. ADS Research will have a focus on: Foundations of Data Science, Health and Life Sciences, Digital Business, City Analytics and Responsible Data Science. Foundations of Data Science
• Data management systems architectures: To facilitate large-‐scale data analysis for statistical data and the analysis of network structures
• Knowledge-‐driven analytics: To use reasoning frameworks and process analytics to study reproducible results for a wide range of data-‐driven problems in business and science
• Machine learning foundations: To determine the quantity of data needed in order to reach concrete conclusions, with a specific focus on health
• Models and techniques for analytics: To address big data variety and veracity challenges • Sustainable systems and software architectures: To build, monitor and analyse complex,
scalable interacting systems for commercial and communication problems Health and Life Sciences
• Bioinformatics: To use data science techniques to study low-‐level molecular and high-‐level behavioural problems in a wide range of applications
• Deep learning and health: To use deep learning techniques in the medical imaging domain through the new AML4Healthcare Research Lab
• Digital health Digital Business
• Big data analytics: To investigate innovative ways for organisations to create real value from (big) data and analytics
• Business model innovation: To find new innovative ways to create and appropriate value • Data Science in business: To bring data science technology to all functions of businesses
(e.g. finance, HR, marketing, supply chain) • IoT: To enable business to get insights from sensor data • Human resources analytics: To enable businesses to improve and innovate through better
insights on talent acquisition, workforce planning, and employee retention City Analytics
• Urban analytics: To apply data science techniques to data from the City of Amsterdam on observations (e.g. traffic streams), interviews and questionnaires
• Urban well-‐being and citizen empowerment: To develop data-‐driven solutions with a predictive and preventive focus on well-‐being as a system of interactions and events
Responsible Data Science • Fairness, Accuracy, Confidentiality, & Transparency (FACT): To future-‐proof responsible
data science methods research is needed focussing on FACT. The methods developed will be inspired by questions from business, government, health, and science
• Ethics: To explore ethics in relation to society and business further • Security: Development of systems to secure privacy
Education: We will investigate innovative teaching and online education possibilities through the Amsterdam School of Data Science. Entrepreneurship: We will collaborate closely with the Amsterdam Center of Entrepreneurship (ACE), e.g. through the Entrepreneurship in Data Science programme (for PhD and Master students), and Innovation Tribe (events to unite academics, corporates and startups). Amsterdam, regional and international network: We will continue to build and expand on our partnerships so that we remain ahead of the field and to ensure financial sustainability from 2019 and beyond. In addition, we will build on our international collaborations, for example through further exchange programmes with Insight Ireland, Tsinghua and Wuhan.
Summary of ADS Events in 2017 (with event coordinators) • 07 December: 2017 Highlights & Looking forward (Maarten de Rijke) • 28 November: AlphaGo Movie Screening (Max Welling) • 10 November: Digital Transformation, Future of Work w/ Coursera (Marc Salomon) • 25 October: Deep Dive into Machine Learning (Max Welling) • 19 October: Responsible Data Science (Maarten de Rijke) • 18 October: Deep Dive into Machine Learning (with Tsinghua University) • 06 October: Coffee & Data on HR Analytics (Sandjai Bhulai) • 08 September: Coffee & Data on Financial Auditing (Edo Roos Lindgreen) • 07 July: Coffee & Data on Visual Analytics (Marcel Worring) • 23 June: Data Science Startup Pitch Event (Maarten de Rijke) • 15 June: Responsible Data Science in Search Engines & Statistics (Frank van Harmelen) • 09 June: Coffee & Data on Controversy in Web Data (Lora Aroyo) • 16 May: ADS Research Assistants Project Highlights (Frank van Harmelen) • 12 May: Coffee & Data on Artificial Intelligence (Maarten de Rijke) • 13 April: Responsible Data Science, Health Genomics & Ethical Society (Frank van Harmelen) • 07 April: Deep Dive into Big Data Management (Stefan Manegold) • 24 March: Launch of Amsterdam School of Data Science (Marc Salomon) • 16 March: Responsible Data Science in Law and Computer Science • 10 March: Coffee & Data on Blockchain (Esther Smit) • 16 February: Responsible Data Science, Software Security & Computational Social Science • 29 January: Responsible Data Science in Law & Statistics (RDS Team) • 13 January: Deep Dive into Machine Learning Deep Dive (Max Welling)
Photos from our events: Launch event of the Amsterdam School of Data Science (March 2017)
ADS visit Tsinghua to discuss collaborative research and staff and student exchange (June 2017)
ADS and Insight Ireland sign agreement for joint research (February 2017)
ADS and Tsinghua sign a memorandum of understanding in Amsterdam (October 2017)