Researcher FED-tWIN project BALaTAI - Belgian Art Links & Tools for Artificial Intelligence

The Royal Institute for Cultural Heritage (KIK-IRPA)

and Ghent University (UGent)
recruit 1 FTE shared between the two institutions for the FED-tWIN project

BALaTAI (Belgian Art Links & Tools for Artificial Intelligence)

Prf-2022-050_BALaTAI

10% Assistant Professor - 40% Postdoctoral research assistant (UGent)

50% SW2 Work Leader (KIK-IRPA)

FUNCTION DESCRIPTION

New computational methods have arguably the potential to transform information inherent to digital art data into a new kind of knowledge, often irretrievable by traditional art-historical methods. Artificial Intelligence (AI) has demonstrated potential to leverage large digital image collections by discovering links among thousands of images and revealing previously undiscovered patterns and connections between different artworks. The invaluable KIK-IRPA’s digital collection BALaT [http://balat.kikirpa.be] with more than 850,000 images (photographs from the 1880s until today) covers all aspects of Belgian cultural heritage. Due to the limitations of the current technologies to extract and visualise information of interest from such huge datasets, it is currently being underused. Hence, KIK-IRPA and Ghent University (UGent) want to recruit a professor / research assistant who will focus on the development of innovative artificial intelligence tools to enable extraction of meaningful insights from this digital dataset and to encourage thereby multiple audiences to engage more with its content.

The needs / tasks of the various end users (KIK-IRPA, heritage researchers, general public, …) must be fully understood, as well as the main methodologies to describe an artwork (and more specifically its composition) and other cultural heritage objects. Attention should also be paid to content retrieval and structure learning as well as knowledge discovery based on AI technologies, and integration with linked open data to enable interactive user experiences. In addition, the translation should be made to related research in the GAIM research group of Ghent University and related GAIM projects should be supported and monitored.

CONTEXT

This function is part of the FED-tWIN program of the Federal Science Policy, which aims to promote sustainable cooperation between the ten Federal Scientific Institution and the 11 Belgian universities through the funding of joint research profiles. The research profile Prf-2022-050_BALaTAI (Belgian Art Links & Tools for Artificial Intelligence) is a joint initiative of the Royal Institute for Cultural Heritage (KIK-IRPA) and Ghent University. Erik Buelinckx (KIK-IRPA) and Prof. Aleksandra Pizurica (UGent) are the promoters.

The position to be filled in consists of a half-time position at the Royal Institute for Cultural Heritage (50% SW2 work leader) and a half-time position at Ghent University (10% professor and 40% postdoc). The candidate must be willing to hold the part-time positions at both institutions and must also apply for both (through a single application, according to the procedure outlined here). Only applying for some and not all of the part-time positions is not possible.

ABOUT KIK-IRPA

The Royal Institute for Cultural Heritage ("Koninklijk Instituut voor het Kunstpatrimonium ‒ KIK" in Dutch and "Institut royal du Patrimoine artistique ‒ IRPA" in French) is a federal scientific institution that takes care of Belgium's heritage: art, monuments and items of historical value. These unique and irreplaceable creations embody our shared history and contribute to our identity as a society. That is why we believe that art generates the emotion that binds us all together. To preserve art and heritage for future generations, we commit ourselves to the highest standards of research and conservation-restoration. This ambition is only possible through a synergy between the diverse disciplines. The Institute brings together art historians, photographers, chemists, physicists, conservators-restorers, experts in image creation, engineers, geologists, etc. Our research gains relevance by bringing together their findings from different angles. In addition, we ensure the transmission of knowledge about works of art, cultural heritage and the technologies of conservation and restoration.

The Information Centre (Infotheek/Infothèque) of the Royal Institute for Cultural Heritage (KIK-IRPA) contains a unique collection of information about Belgium’s artistic and cultural heritage. It houses both a photo library (his photo archive is Belgium’s visual memory, which contains more than 1 million photos of movable and immovable heritage, linked to over 460,000 descriptions of cultural heritage artefacts and representations) and library (with works dedicated to art history, the conservation and restoration of works of art, and the chemical, physical and material research linked to art restoration), and it manages the institute’s intervention files and archives (more than 20,000 files of advice on art history, preliminary studies and conservation-restoration files). The continuing process of digitization feeds daily the freely accessible online BALaT database [https://balat.kikirpa.be].

ABOUT GHENT UNIVERSITY

Ghent University [https://www.ugent.be/] is one of the most important educational and research institutions in the Low Countries. More than 9,000 employees and 41,000 students live up to the credo “Dare to Think” every day. High-quality education, internationally acclaimed research and a pluralistic social responsibility characterise the mission of Ghent University.

The research group GAIM (Group for Artificial Intelligence and Sparse Modelling) of the Faculty of Engineering and Architecture, Department Telecommunications and Information Processing will collaborate with the KIK-IRPA for the proposed FED-tWIN project. GAIM [https://gaim.ugent.be] conducts fundamental and applied research in various domains relevant to this vacancy (such as analysis of multimodal images, data mining, deep learning, clustering and data fusion). The available technical expertise in the development of building blocks for the enrichment of (heritage) collections is of great importance to start and support the FED-tWIN project.

PLACE OF EMPLOYMENT

• Royal Institute for Cultural Heritage and Ghent University, Jubelpark 1, 1000 Brussels

• Ghent University, Faculty of Engineering and Architecture, Department of Telecommunications and Information Processing TELIN (Group for Artificial Intelligence and Sparse Modelling – GAIM):  Sint-Pietersnieuwstraat 41, 9000 Ghent

JOB DESCRIPTION

The first phase of the FED-tWIN employment mainly focuses on a pilot project that will analyse and test the latest AI-tools on preselected subcollections from BALaT and develop novel methods in response to the defined research objectives and technical challenges. Particularly, in the first two years the focus will be on developing (1) a deep active learning (DAL) framework for detecting objects and features of interest and (2) an unsupervised deep learning approach for image retrieval and identifying related works. In this timeframe, the professor-researcher will also investigate, in close collaboration with KIK-IRPA staff and management, how these tools can be implemented in a new version of the BALaT-platform that is currently under development. These actions will help increase the interoperability of the collection data and facilitate its digital dissemination to a wide audience.

Within the first five years of the project, the goal is also to leverage the collection by enriching and linking the data, e.g., to enable visual links and to integrate semantic links from other open-source datasets.  With the support of KIK-IRPA’s art historians and documentalists it will be sought to enhance user-engagement through human-centred, interactive and pedagogic design of the user-interface. The objective is to enhance knowledge discovery from images, contextual information from metadata and unstructured data stored in BALaT. The developed tools should also enable non-expert end-users to obtain easily rich overviews of the collection based on a given query, as well as relevant relationships between collection features.

In the second research phase (after 5 years) the scope will be expanded towards fully implementing innovative digital technologies in the working processes of KIK-IRPA, improved data enrichment and linking and more advanced search strategies for researchers. Next to further development of the AI tools, attention will be paid to the development of a digital storytelling platform, where online visitors can create their own narratives using photographs, videos, audio, text and images with computer-based tools. Such a platform should encourage online visitors to explore, to reuse and personally connect to online cultural heritage collections.

Next to his/her personal research on the project, the selected candidate will guide doctoral students, gradually build a research team, attract funding to expand this research and take part in teaching activities in the domain of this project.

PROFILE

The candidate is expected to demonstrate mastery of computer vision/image processing, machine learning and/or data science. Experience with multi-modal and/or high-dimensional data processing, and particularly with pattern recognition, clustering and content retrieval is an asset as well as knowledge and experience with deep learning architectures.

Strong mathematical background and the ability to perform well-founded and technically sound research are crucial. Furthermore, the candidate must demonstrate the ability to express the research goals, vision and scientific contributions in a technically sound and clear manner, which should be evident from her/his publications.

The candidate must demonstrate the affinity and the ability to work within an interdisciplinary team. The affinity to cultural heritage is important, and prior experience with processing cultural heritage data is an asset.

The candidate must demonstrate the affinity and the ability to guide PhD students, attract funding and develop a research team.

The candidate is able to combine and validate academic and practical knowledge and is able to communicate about it in an understandable way for the end users. Knowledge of the most important Belgian legislation and European guidelines on copyright and reuse (GDPR) in the heritage sector is recommended. Furthermore, he/she is aware of the objectives of Open Science and can find his/her way through datasets and tools/platforms relevant to the field.

S/he must actively participate in profiling Ghent University, GAIM and the Royal Institute for Cultural Heritage as leading figures within the cultural heritage domain and the GLAM sector (Galleries, Libraries, Archives and Museums) in the field of innovative enrichment techniques applied to cultural heritage.

The candidate must also consider the development of a data management plan (in accordance with the FAIR research data management principles and as well BELSPO as UGent DMP recommendations), ensure coordination between the work packages, and stimulate collaboration with colleagues in both organizations. The candidate must also comply with the ’Code of Ethics for Scientific Research in Belgium‘, which sets out the main principles of ethically responsible scientific practice.

KNOWLEDGE/EXPERTISE

• Experience in conducting scientific research at top level and writing high-quality publications.
• At least two recent publications in top technical journals in the field of this project, and recent presentations at international conferences in the field related to this proposal.
• Lead (or participated) in writing project proposals to acquire funding in one of the above-mentioned domains.
• Experience in supervising doctorates and master theses.
• Considered as a plus: the candidate already worked with heritage data (e.g., collection registration and access) or has relevant experience within a cultural heritage institution.
• Recommended is: having experience in international mobility, including through participation in research programs at research institutions not affiliated with the university where the highest degree was obtained.

More specific skills:

• The candidate has a strong theoretical background in basic maths and probabilistic reasoning, signal/image processing and machine learning, and can demonstrate strong analytical and problem-solving capacity.
• The candidate also has excellent writing skills, is able to present the research ideas and research results in a clear and technically sound way and is in general able to articulate well the main ideas, novelties and the significance of her/his work.
• The candidate must be motivated to work in the cultural heritage domain and to collaborate within an interdisciplinary team.
• Self-motivation and enthusiasm for research, eagerness to learn and contribute to the research done within the group, are crucial. The candidate is a team player and takes initiative where needed.
• Experience in organizing and leading meetings is required. The candidate is willing and able to guide junior researchers (PhD students), mentor them in their research and writing papers.
• Programming skills are important (e.g., in Python, Matlab, C++/C#, JavaScript, …). Furthermore, any experience with data visualization and/or user interfaces design is a plus.
• The candidate has a good knowledge of English (both orally and written) has good communication skills, flexibility and critical spirit to collaborate with researchers of various nationalities and disciplines.

CONDITIONS FOR ADMISSION

The intended candidate holds a degree of Doctor of Computer Science Engineering or Doctor of Computer Science or an equivalent degree.

When assessing a foreign (non-EU) diploma, an equivalence certificate may still have to be requested from NARIC; we recommend that you - if necessary - start the recognition procedure at NARIC as soon as possible. You must have this recognition no later than the date of appointment.

The PhD degree is obtained at the earliest 12 years prior to the submission of the job application with derogations. The 12-year period is extended by one year for each maternity, parental or adoption leave of the candidate and for each long-term sick leave of the candidate or his/her immediate family.

APPOINTMENT INFORMATION

The position consists of a joint commitment of both partners:

• A half-time employment contract (50%) for an indefinite period (19 hours per week) as a SW2 work leader), at the Royal Institute for Cultural Heritage: Jubelpark 1, 1000 Brussels;
• An appointment at Ghent University as a part-time professor (10% assistant professor (docent) for an initial period of 5 years, renewable) and as a postdoc researcher (40%, indefinite duration, scale PD1 to PD4), Department of Telecommunications and Information Processing, Sint-Pietersnieuwstraat 41, 9000 Ghent.

The candidate must be willing to hold the part-time positions at both institutions and must also apply for the above-described joint commitment (through a single application via e-recruitment). Only applying for some of the part-time positions listed above is not possible. The candidate adheres to the regulations of both institutions.

Date of employment: February 1, 2024.

OFFER

KIK-IRPA - benefits

As a staff member of the KIK-IRPA you can count on a number of benefits such as a bicycle allowance, free commuting by public transport, free hospitalization insurance, all kinds of social benefits, flexible working hours and a workplace in a beautiful location and easily accessible by public transport.

• A contract of employment at the grade of SW2 work leader at salary scale SW21 or SW22, on a half-time (50%) indefinite period.
• Minimum remuneration (gross amount at current index, excluding statutory allowances): SW21: €28.873,71 per annum.

Additional benefits:

• Possibility to benefit from a bilingualism allowance;
• Extensive training offer (to be taken during working hours);
• Easy access via public transport;
• Possibility of receiving a cycling bonus;
• Interesting offers and benefits with the FED+ card;
• An ambitious and friendly team.

Ghent University - benefits

The independent academic staff (professor) career policy is based on talent development and growth, in which vision development and strategy - both personal and at group level - are central. Ghent University focuses on career guidance and coaching of the faculty member in the various phases of the career.

More information about the career progression of a professor (in Dutch):

https://www.ugent.be/zaploopbaan 
 

Every UGent staff member can count on a number of benefits such as a cycle kilometer allowance, bicycle rent/repair, reimbursement of public transport commuting, childcare, sports facilities with free insurance, teleworking, eco vouchers, ...

https://www.ugent.be/en/work/talent/next-steps

HOW TO APPLY?

You can only apply online via e-recruitment before the application deadline (see above). We do not accept late or incomplete applications, or applications that are not sent through the online system.

Your application must include the following documents:

• In the field ‘Application form’: the professorial staff application form (+ all annexes mentioned in the form), merged into one PDF file.
• In the field ‘Cover letter’: your application letter in PDF format
• In the field ‘Diploma’: a transcript of your doctoral degree. If you have a foreign diploma in a language other than our national languages (Dutch, French or German) or English, please add a translation in one of the mentioned languages.
• In the field ‘Certificate of equivalence’: only for diplomas awarded outside the European Union: certificate of equivalence (NARIC) (if already in your possession). For diplomas awarded in the UK before January 31st of 2020, a certificate of equivalence is not required.

Note that the maximum file size for each field is 10 MB.

KIK-IRPA and Ghent University pursue an equal opportunities and diversity policy and therefore encourage everyone to apply.

ADDITIONAL INFORMATION

For additional information about the project and the position to be filled, please contact the promoters: Erik Buelinckx – erik.buelinckx@kikirpa.be, and Prof. Aleksandra Pizurica – aleksandra.pizurica@ugent.be.

For additional information about the selection procedure, please contact: vacature@kikirpa.be (KIK-IRPA) or the Recruitment and Selection department of Ghent University (selecties@UGent.be).

Additional information about the FED-tWIN research programme: FED-tWIN | Research Programmes | Research Programmes and Infrastructures | Belspo

Vragen over online solliciteren? Bekijk de FAQ of contacteer de helpdesk (dpo@ugent.be - tel. 09 264 98 98)