The Wide Role of Informatics at Universities

THE WIDE ROLE OF INFORMATICS AT UNIVERSITIES Elisabetta Di Nitto Susan Eisenbach Inmaculada García Fernández Eduard Gröller The Wide Role of Informatics at Universities RESEARCH EVALUATION of Computer Science (Draft) An Informatics Europe Report Report Title Prepared by: •Elisabetta Di Nitto, Politecnico di Milano, Italy •Susan Eisenbach, Imperial College London, United Kingdom •Inmaculada García Fernández, University of Malaga, Spain •Eduard Gröller, TU Wien, Austria The Wide Role of Informatics at Universities October 2019 Published by: Informatics Europe Binzmühlestrasse 14/54 8050 Zurich, Switzerland www.informatics-europe.org [email protected] © Informatics Europe, 2019 Other Informatics Europe Reports ●Industry Funding for Academic Research in Informatics in Europe. Pilot Study. (2018, Data Collection and Reporting Working Group of Informatics Europe). ●Informatics Education in Europe: Institutions, Degrees, Students, Positions, Salaries. Key Data 2012-2017 (2018, Svetlana Tikhonenko, Cristina Pereira). ●Informatics Research Evaluation (2018, Research Evaluation Working Group of Informatics Europe). ●Informatics for All: The strategy (2018, Informatics Europe & ACM Europe) ●When Computers Decide: Recommendations on Machine-Learned Automated Decision Making (2018, Informatics Europe & EUACM, joint report with ACM Europe) ●Informatics Education in Europe: Institutions, Degrees, Students, Positions, Salaries. Key Data 2011-2016 (2017, Cristina Pereira, Svetlana Tikhonenko). ●Informatics Education in Europe: Are We All In The Same Boat? (2017, The Committee on European Computing Education. Joint report with ACM Europe). ●Informatics in the Future: Proceedings of the 11th European Computer Science Summit (ECSS 2015), Vienna, October 2015 (2017, eds. Hannes Werthner and Frank van Harmelen, Springer Open). ●Informatics Education in Europe: Institutions, Degrees, Students, Positions, Salaries. Key Data 2010-2015 (2016, Cristina Pereira). ●Informatics Education in Europe: Institutions, Degrees, Students, Positions, Salaries. Key Data 2009-2014 (2015, Cristina Pereira). ●Informatics Education in Europe: Institutions, Degrees, Students, Positions, Salaries. Key Data 2008-2013 (2014, Cristina Pereira, Bertrand Meyer, Enrico Nardelli, Hannes Werthner). ●Informatics Education in Europe: Institutions, Degrees, Students, Positions, Salaries. Key Data 2008-2012 (2013, Cristina Pereira and Bertrand Meyer). ●Informatics Doctorates in Europe - Some Facts and Figures (2013, ed. Manfred Nagl). ●Informatics Education in Europe: Europe Cannot Afford to Miss the Boat (2013, ed. Walter Gander, Joint report with ACM Europe) All these reports and others can be downloaded at: www.informatics-europe.org E S In this report, we discuss the results of the online survey conducted by Informatics Europe Working Group on the Wide Role of Informatics at Universities. The main goals were to understand the value universities place on interdisciplinary research and teaching, what happens in practice with hiring and supporting interdisciplinary academics, and what structures are in place to support interdisciplinary work. We also examined Data Science’s impact in detail, given its rapid rise and importance. Forty eight universities from nineteen European countries have participated in the survey providing answers on these strategic topics. The results of our invesgaon have shown that: •In any area examined a significant majority of surveyed universies were engaged with interdisciplinarity. However, there were Informacs academics concerned about the development of interdisciplinary research mainly owing to limited funding, low esteem compared with discipline-specific research or lack of strategic direcon. •The majority of surveyed universies run joint degrees, including Informacs, most frequently in the area of Business and Economics, Natural and Life Sciences, and Engineering. The universies not already offering joint degrees showed a considerable interest in running new joint degrees including Informacs in Mathemacs and Stascs, Natural and Life Sciences, Law, Social and Polical Sciences, and Business and Economics. •With regard to teaching of Informacs in non-informacs programmes, there was not a uniform paern. While for some universies there existed a clear disciplineresponsibility, in others there was no clear policy about which department teaches Informacs in non-informacs programmes. Moreover, in several universies the lack of human resources prevented the Informacs departments from being in charge of teaching Informacs subjects in non-informacs degree programmes. •In terms of policies for interdisciplinarity and financial support for staff and centres, the range of answers was very large from no policy or financial support to using significant resources for hiring staff and seng up and funding centres. In the case where universies were largely autonomous from naonal agencies, hiring interdisciplinary researchers was encouraged when there was some funding, oen by third pares, dedicated to this. Respondents from countries where the hiring system was strongly regulated by some naonal agency highlighted the difficulty to introduce some flexibility and to define long-term plans which include muldisciplinarity. •The most commonly found centres were in Data Science, an area which was largely seen to emerge from Informacs and Stascs. According to the majority of surveyed universies, the rise of Data Science has changed the percepon of Informacs resulng in increasing relevance of ethics and other social aspects and in developing introductory courses on digital literacy and skills in all study programs. Informacs was considered to be the main knowledge centre in the digital transformaon of society and many ini- aves are under way changing how Informacs is perceived. For all the quesons there were a significant number of universies that have not engaged in official interdisciplinary acvity. 1 1. I In the 1970s with the advent of the personal computer we entered into the Digital or Informaon Age. However it has only been in this century with the ubiquity of the internet, the smartphone, cloud, and the internet of things that digital has become truly pervasive. How do universies respond to this massive change? Informacs Europe established in 2018 a new working group to invesgate what universies are doing to ensure that non-informacs teaching and research is informed by best pracce in Informacs. To beer understand the state of affairs on this topic and discover best pracces at European Universies, the working group conducted an online survey. We invited heads and members of Informacs/Computer Science/IT Departments (Schools, Facules, Instutes) to complete a quesonnaire in autumn 2018. The request to fill out our survey was sent to all Informacs Europe members and it was also publicly available from the Informacs Europe website. For the locaon of the respondents see Figure 1. Forty eight universies from nineteen countries filled it out (see Appendix B). Our survey was wide ranging. We wanted to understand how universies valued interdisciplinary research, about teaching Informacs to non-specialist students, what happens in pracce with hiring and supporng interdisciplinary academics, and what structures are in place to support interdisciplinary work. We chose to examine Data Science’s impact in detail, given its importance and newness. For the actual survey quesons see Appendix A. Although how Informacs (also called Computer Science or Compung) should posion itself in a university is a polical decision, in many universies what happens has arisen organically rather than strategically. There are a wide range of models with the extremes ranging from primarily being a service department to being primarily a research area that is isolated from other departments. 2.  Universies are normally structured into disciplines which foster disciplinary research. However, the ubiquity of Informacs in our culture has led to pressures for research that is interdisciplinary. Pressures in favour of such research comes from academics themselves, student interests, external funding sources, and somemes from university leadership. The following subsecons discuss the answers obtained for each specific queson. 2 3 () Countries () Regions F 1. Locaon of Respondents F 2. What is the University atude towards Interdisciplinary research? 2.1. Desirability of interdisciplinary research. The first part of the survey quesoned respondents on university atudes and acons in respect of interdisciplinary research.1 A large majority (71%) claimed that their university encouraged interdisciplinary research when compared with single discipline research (see Figure 2). This seems to imply that universies favour interdisciplinary research over single discipline research. However, several respondents indicated that their encouragement was largely ‘theorecal’ and accompanied by lile, if any, funding. Some respondents said that much of the interdisciplinary work at their instuon occurred between departments other than Informacs. Only one respondent indicated that their 1The survey does not differenate between interdisciplinary work in general and that with an Informacs component. Given who answered the quesonnaire, one can assume that Informacs is included. 4 university actually discouraged interdisciplinary research although others menoned that their departments were judged, usually naonally, against discipline-specific criteria. F 3. What is the Department atude towards interdisciplinary research? 2.2. Department atude towards Interdisciplinary research. With the same queson directed at Informacs Departments rather than the whole university (see Figure 3), two thirds of respondents sll claimed that interdisciplinary research was favoured over single discipline topics. However, similar comments are made about encouragement being in principle rather than in pracce and about being judged on discipline-specific criteria. F 4. Are there interdisciplinary areas of research where your university could enter but aren’t due to lack of university support? 2.3. University support. However, just over half (51%) of the respondents recorded (see Figure 4) that their university supported all areas of interdisciplinary research which required support. Others (30%) menoned a variety of potenal Informacs areas where university support for interdisciplinary research was lacking. Others talked of the need for strategic planning to direct interdisciplinary efforts or of the need to focus given the wide range of potenal opportunies. 5 F 5. Are there other players who have helped increase the interdisciplinary research in your university? 2.4. Addional support. When asked about external support for interdisciplinary research directed towards their university (see Figure 5), 50% of the respondents responded posively with 40% stang that naonal public funding sources had helped to increase interdisciplinary research. A further 22% mainly discussed specific formal or informal arrangements between their department and others in their instuon. 2.5. Final thoughts. Respondents were asked to make some more general comments. Not all respondents were especially supporve of interdisciplinary research per se. It was noted that, because some funding streams demand interdisciplinarity, it is possible that ‘arficial collabora- ons’ were formed that aracted the funds but did not make good use of the capabilies of the researchers. Frequently interdisciplinary projects are focussed on how informaon technology can serve the other discipline so the progress made and any breakthroughs that occur advance the other discipline but have no impact on the development of Informacs. One respondent suggested that the excitement and interest in supporng interdisciplinary projects could make it likely that lower quality proposal were accepted (compared with single discipline ones). Respondents with more posive atudes towards interdisciplinary research were oen nevertheless concerned about its development mainly owing to limited funding, low esteem compared with discipline-specific research or lack of strategic direcon. 3.  When teaching is run by departments it is easier to have single discipline degrees rather than joint degrees, and there is no shortage of students wanng to study Informacs as a single discipline. Nonetheless there is pressure (from prospecve students, academics, industry, and 6 somemes university leadership) to have joint degrees. The following subsecons discuss the answers obtained for each specific queson. F 6. Does your university run joint degrees? 3.1. Joint degrees. 30% of the universies do not run a joint degree that includes Informacs (see Figure 6). Within this group of universies, some specified that all their programs entail technical aspects of IT, such as programming or data base technology. At some of these universies there are plans for some joint programmes, e.g. a Data Science BSc programme that joins Computer Science, Maths and Industrial Engineering, and an MSc in Game Design and Produc- on jointly with the Arts School. These are collaborave iniaves in new direcons, where the Informacs Department is one of the partners. Occasionally another department has a small Informacs group who provides the Informacs teaching for a joint subject degree. The remaining 70% of the universies run joint degrees, the most popular joint degrees including Informacs are Business and Economics (Business Informacs; CS and Business; Compung and Economics; Informaon Systems combining Informacs and Business Administra- on; CS and Management; Informacs and Economics; Informacs and Finance; Economics and Business Informacs; Data Science and Entrepreneurship) followed by Mathemacs and Stascs (Informacs and Mathemacs; Data Science; Informacs and Applied Mathemacs; Informacs and Stascs), Natural and Life Sciences (Bioinformacs; Informacs and Natural Sciences; CS and Physics; AI for Biomedicine; Precision Medicine; Geoinformacs; Chemistry and Informacs; Biology and Informacs; Informacs Health) and Engineering (Computaonal Engineering; Computer Engineering; Electronics and Informaon Engineering; Informacs and Electronics; Informacs and Telecommunicaons; Informacs and Cybernecs; Informacs and Mechatronics; Informacs and Aerospace Engineering; Informacs and Civil Engineering; Informacs and Industrial Engineering). Joint degrees in Informacs plus Arts, Design and Media (Technical Communicaon; Design Informacs; CS and Communicaon, CS and Design; ICT and Media; Informacs and Informaon Science; Informacs and Library Science) or Law, 13 the scientific and societal impact of Data Science and Arficial Intelligence rather in the (external) applicaon domains, although an increase in Informacs students is recognisable. The early awareness of Data Science and Machine Learning as areas of rapidly increasing importance is considered crucial. Due to ineral forces (especially at larger universies), however, somemes acve strategies from the top university level are lagging, though boom up approaches might compensate for this. As in analogous situaons in the past, Informacs is struggling to be viewed only as a service department to help other domains in solving their Data Science problems. This is similar to previous interdisciplinary approaches (e.g., Mulmedia, Computer Graphics, Animaon) where Informacs is used as a tool, but gradually also as a research partner on an equal foong. The surge in interest in Data Science is accompanied by larger resource flows. The uncertainty about where to locate the Data Science acvies might lead to the simultaneous development of several research groups at one university. This decentralised approach might allow the different departments to grow and manage their own Data Science groups with discriminave strengths. The quickly amplified interest in Data Science is primarily considered an opportunity, where it is challenging to follow and sustain all parallel acvies. Currently the interest in Data Science, Machine Learning, and Arficial Intelligence is so large that this might overshadow all other areas of Informacs. Too imbalanced funding opportunies and student flows should be avoided to provide a well-adjusted porolio of competences to the society and economy. 5.4. Final Thoughts. The interest and popularity of Data Science and Arficial Intelligence has dramacally risen in the last 10-20 years. These technologies have the potenal to be driving and enabling technologies for the rapidly unfolding digital transformaon of society. The very fast developments lead to many daunng challenges, e.g., concerning privacy, security, bias, reliability, robustness, legal and ethical implicaons. It is not yet clear where Data Science should be anchored, e.g., in the Informacs Department, mul-department units, applicaon domains, also. Due to the developmental speed, established organisaons like universies are struggling to swily adjust their organisaonal structures and educaonal porolios, where long term changes have yet to be implemented. For some experts in potenal applicaons fields Data Science and Arficial Intelligence might be perceived as a hype that will cool down eventually. Despite this, most experts see the growing and pervasive importance of Informacs methods for their research area. The Data Scienst as a profession will be much more heterogeneous in the required skill set as compared to other interdisciplinary approaches, like Business Informacs, Bio-Informacs, or Medical Informacs, which basically involve two disciplines each. Considering the wide array of concerned fields, the Data Scienst will have a deep knowledge in just one or a few speciales and have a broad (and shallow) knowledge of the many other concerned areas. Data Science encompasses a mixture of muldisciplinary skills ranging from Mathemacs/Stascs, Programming/Databases, Domain Knowledge/So Skills, Communicaon and Visualisaon. The fluidity of the development and the breadth of the area will transfer to Data Science groups, centres, and curricula with largely varying specialisaons. It seems very likely that Informacs will play a key role in all these developments, where we should proacvely use the many emerging opportunies. 14 6.  Creang actual rather than virtual interdisciplinary centres is likely to improve the chances of interdisciplinary research and teaching lasng. The following subsecons discuss the answers obtained for each specific queson. F 16. Does your university set up centres for interdisciplinary work? 6.1. Interdisciplinary centres. 28% of respondents said their university did not have real interdisciplinary centres (see Figure 16). Of those who commented on why there was a lack of centres only one actually replied that their management was averse to seng up addional administrave structures. The rest just said there were informal groupings, but nothing officially supported. 45% of all of the interdisciplinary centres were set up primarily for research and only 18% for teaching. The rest were primarily involved with industry. There were a broad range of centres in the different universies – clearly what experse is in a university and what the structure of the different departments/schools/facules impacts which centres are set up in addion to the exisng primary structures. The most common centres menoned with a significant Informacs component were in Computaonal Science, Data Science, Life Science, Digital Society, Energy, and Security. There were also more than one university with the following centres: Biomedical Engineering, Environment/Climate, Medical Imaging, and Complex Systems. There were a wide range of centres which only menoned at one university: Health, FinTech, Digital Humanies, Roboc Surgery, Cognive Ageing, Bioinformatics, and Geoinformacs. F 17. Why were the centres created? 15 6.2. Purpose of interdisiciplinary centres. 45% of all of the interdisciplinary centres were set up primarily for research and only 18% for teaching (see Figure 17). The rest were primarily involved with industry collaboraon or consultancy. F 18. Which enty control the interdisciplinary centres? 6.3. Ownership of interdisciplinary centres. Of the 36 respondents, 21 (or 58%) were independent enes within their university, 12 (or 1/3) were co-owned by the departments that are involved and the rest had a single department that owned them (see Figure 18). It is surprising that so many were separate enes as this means if they are not self-funding money will be an issue. F 19. Where are the centres located? 6.4. Locaon of interdisciplinary centres. More than half of the respondents said that the centres they were reporng on were located ‘elsewhere’ on campus (see Figure 19). Although a significant minority described the centres as ‘virtual’ implying that they actually had no physical locaon. One contributor disnguished between a large centre that had its own space, and smaller ones that were embedded in departments. Others spoke of large buildings that accommodated many different groups such that a nearby centre may not be associated with a department. 6.5. Funding of interdisciplinary centres. Only 25% of the interdisciplinary centres reported on were funded enrely externally, the funding sources of the rest were equally split between enrely internal and mixed sources of funding (see Figure 20). In the majority of cases where funding is enrely internal, the bulk of the actual cash seems to come from central funds with departments providing resources ‘in kind’. Frequently, me-limits were expressed (five and six years are menoned) aer which the centre is expected to be self-financing. For the universies 16 F 20. Who funds interdisciplinary centres? that reported on (enrely or parally) external funding, in many cases only government and EU programmes were explicitly cited as sources of funds. F 21. Are there changes planned for seng up or closing centres? 6.6. Planning for changing interdisciplinary centres. A quarter of respondents reported on plans to set up new centres (see Figure 21). Some described a noon of connuous evoluon of interdisciplinary work. Only AI was explicitly menoned as a target for the development of new centres. Other respondents, although not explicitly planning a new centre, menoned the issue of the periodic review of exisng centres cing various opons including merging centres and/or creang new centres. 6.7. Drivers for new acvies. Nearly one third of respondents reported on internal drivers and pressures bearing on innovave acvity (see Figure 22). Amongst the drivers, academic curiosity of staff and students was cited alongside a need for research collaboraon. Pressures included demands to increase students enrolment, to modify the curriculum and university ini- aves to set up a centre. One university also menoned limitaons of student numbers and limitaons on joint degrees that inhibited their development goals. The other respondents addressed external drivers and pressures. The most significant cited pressure concerned the societal influence of globalisaon together with an associated driver on universies to promote innovaon and technology transfer (47%). The next most significant pressure was the search for funding driven by government iniaves (30%) whilst other respondents observed the expanding role of Informacs in other disciplines and the pressure on Informacs Departments to support these disciplines (20%). Finally, one respondent menoned compeon between universies as an external pressure. 17 F 22. What are the drivers and pressures for new centres? F 23. Is any support provided for interdisciplinary work? 6.8. Support for interdisciplinary work. Respondents were evenly split over this queson (see Figure 23) although several of those who claimed instuonal support were rather equivocal - ”I would guess so” and “Some departments …”. Respondents who reported no instuonal support divided into those who spulated some form of external support and those who did it “as a hobby” ( 25%). F 24. Are there any centres for interdisciplinary work created from strategic iniaves? 18 6.9. Strategic vision. More than half of the respondents reported on centres created from strategic iniaves (see Figure 24). Many of these were oriented towards Informacs themes (FinTech, Crypto-currencies, Data Science) but several other types of centre were menoned (Learning and Educaon, Cultural Heritage, Sustainability and Energy). F 25. Is there an official strategy to widen the role of Informacs? 6.10. Official strategic vision. Respondents were exactly split on this queson (see Figure 25). Of those who answered posively, the emphasis was on muldisciplinarity for about half the respondents. Informacs topics cited by others included Cyber Security, Data-driven Innova- on, Intelligent Systems, Applied Computer Science and Digital Humanies. Respondents who answered “No” were not very forthcoming with their comments. 6.11. Final thoughts. Nineteen respondents contributed their overall views on the current situaon in their universies. One response was wholeheartedly supporve cing good funding, strong collaboraon and a sound internaonal reputaon as aracve to world-class researchers. Other respondents menoned limited or non-existent funding and other, higher priories (like increased student enrolment as factors which retarded interdisciplinary inia- ves. Two universies thought that Informacs was too junior a partner in the context of their university to make much impact. By far the most significant issue concerned the nature of either the central or departmental strategic direcon. Three respondents asked for greater freedom for individual researchers to be more creave with ideas, contacts and funding. However, there were ten contributors who asked for beer communicaon between facules, more structured research management or further internaonalisaon. A few just wanted more substance to the strategy - “It is only a goal without supporng instruments. ”; “Sll under construcon - too early to conclude …”. 7. C Despite the ubiquity of Informacs, in any area we examined there were a significant minority of surveyed universies that have not really engaged with interdisciplinarity. This does not preclude individual academics within these universies working on muldiscipline research and teaching. On the other hand there are Informacs academics who are concerned that the pressure towards muldisciplinary research is at the cost of core Informacs research. How much a university’s leadership want to encourage interdisciplinarity can be seen in its policies and financial support for staff and centres. The range is very large from no policy or financial support to using significant resources for hiring staff and seng up and funding centres. The most commonly found centres are in Data Science and this is an arena which is largely seen to 19 arise from Informacs and Stascs. It seems quite early to see a paern on how universies are going to develop with respect to interdisciplinary research. As to joint teaching, there are a very wide range of courses offered that include Informacs. Acknowledgements. We would like to thank all the respondents who wrote many thoughul answers and provided the raw data. In addion to the authors, there are many people who have put significant me into both designing the quesons and reading earlier dras. These include Stuart Anderson, Luis Caires, Brian Keegan, Ulf Lesser, Chris Sadler, and Hannes Werthner. Finally, this document would not exist without the broad and extensive help from Svetlana Tikhonenko of the Informacs Europe office. 20 A A. S: T W R  I  U (1) Research (a) When compared with single disciplinary research, does your university encourage or discourage (or neither) interdisciplinary research? If so how? (e.g. funding, me, physical centres) •Encourage •Discourage •Neither encourage nor discourage (b) Does your Informacs Department encourage or discourage (or neither) interdisciplinary research? If so how? •Encourage •Discourage •Neither encourage nor discourage (c) Are there interdisciplinary areas of research where your university could (should) enter but aren’t due to lack of university support? If so what are they? (d) Are there other players who have helped increase the interdisciplinary research in your university? For example has a funding body focused a programme on interdisciplinary PhD studentships which academics applied for?If so what external organisaons and what programmes have increased interdisciplinary research at your university? (e) Please comment on any advantages or disadvantages you perceive of your university’s arrangements. (2) Teaching (a) Does your university run joint degrees (e.g. X and Informacs, Informacs and X, X with Informacs, Informacs with X). If yes, what are they? •Yes •No (b) Are there plans to run new joint degrees or to close down joint degrees? If yes what are they? •Run new joint degrees •Close down joint degrees •Neither run nor close down (c) Who teaches the Informacs component of non-informacs degrees? For example, is programming taught to Physicists by members of the Physics Department, of the 21 Informacs Department or is there a servicing organisaon within your university that teaches Physics students to code (or some other mechanism)? (d) If Informacs is taught by people not located in an Informacs department are they Computer Sciensts by training or research? •They are Computer Sciensts •They are not Computer Sciensts •Informacs is not taught by people not located in an Informacs Department (e) Please comment on any advantages or disadvantages you perceive of your university's arrangements. (3) People (a) Does your university explicitly adverse/hire academics who focus on interdisciplinary research? •Yes •No (b) Are they rooted in a department, have a joint appointment across departments, or rooted in a centre? •Rooted in a department •Have a joint appointment across departments •Rooted in a centre (c) How is their quality judged for both appointment and for promoon?For example are they judged according to the criteria of one of the departments or both? Are the people who judge from a single department or both? (d) Are there any iniaves planned to hire in interdisciplinary areas? •Yes •No (e) Please comment on any advantages or disadvantages you perceive of your university's arrangements. (4) Data Science (a) Which department in your university is seen to own this area? Is it Informacs, Stascs, jointly or somewhere else? •Informacs Department •Stascs Department •Jointly Informacs and Stascs Department 22 •Somewhere else (please specify) (b) Has the rise of this area changed the percepon of Informacs overall in your university? •Yes •No (c) Please comment on any advantages or disadvantages you perceive of your university's arrangements. (5) Structure (a) Does your university set up centres for interdisciplinary work? If yes can you say which they are? •Yes •No (b) Are they for research, translaon (technology transfer), consultancy, and/or teaching? •Research •Translaon (technology transfer) •Consultancy •Teaching (c) Are they rooted in a single department (say which one), owned by the departments involved or independent? •Rooted in a single department •Owned by the departments involved •Independent (d) Are they physically located within a department, nearby or elsewhere on campus? •Within a department •Nearby a department •Elsewhere on campus (e) How are any centres funded? Does the university provide any money to startup or are they funded by external money? Does the university provide longer term money? (f) Are there plans to set up more centres or to close centres? If so what will they be? •Set up more centres •Close centres