Thesis: BA/MA

Background: 

Graphs are a natural representation in a wide range of real-world applications, such as social networks, biological interactions, linguistics, and computer networks. Graph analysis can provide a deeper understanding of such systems and help to predict their behavior. A number of techniques have been developed to address tasks such as node classification, node recommendation, link prediction, and anomaly detection. More recently, techniques to transform graphs into a vector space representation, i.e., graph embeddings, have opened up new opportunities to tackle these tasks. 

Research problem: 

Representing log data as graphs (e.g., network traffic, computer events, software behavior) allows an analyst to easier navigate and explore the events and their context. Spotting undesired, often rare events inside graphs still requires a lot of manual work. Using word embedding techniques for anomaly detection in log data is a new field, which is in the focus of this thesis. You will: i) study existing word embedding approaches for log data (e.g., IP2Vec, Log2Vec), ii) reproduce and explore those approaches, iii) find interesting new application areas and datasets, and iv) extend existing embedding approaches if necessary and evaluate them on the new datasets.

Three initial references:

1. M. Ring, A. Dallmann, D. Landes and A. Hotho, "IP2Vec: Learning Similarities Between IP Addresses," 2017 IEEE International Conference on Data Mining Workshops (ICDMW), New Orleans, LA, 2017, pp. 657-666, doi: 10.1109/ICDMW.2017.93.

2. Fucheng Liu, Yu Wen, Dongxue Zhang, Xihe Jiang, Xinyu Xing, and Dan Meng. 2019. Log2vec: A Heterogeneous Graph Embedding Based Approach for Detecting Cyber Threats within Enterprise. In Proceedings of the 2019 ACM SIGSAC Conference on Computer and Communications Security (CCS '19). Association for Computing Machinery, New York, NY, USA, 1777–1794.

3. Tuulio, V 2020, “Review of popular word embedding models for event log anomaly detection purposes.” Master Thesis, University of Helsinki

 

Thesis:BA

Background: 

Cybersecurity has become a key issue in modern information economies. Security incidents, such as intrusions, data breaches, denial of service attacks, and ransomware infections are routinely reported in the media and illustrate that many organizations struggle to cope with a growing variety and intensity of threats. The business impact of security incidents -- operational, financial, legal, and in terms of long-term losses of consumer confidence and trust -- can be significant. 

Security management aims to assess and control such information security risks based on known and newly identified vulnerabilities, weaknesses, and threats. Understanding the latter, i.e., adversaries and their intentions, goals, tactics, and techniques, is fundamental for risk assessment, selection of effective countermeasures, classification of attacks and for decisions on how to respond to them. To this end, various standards and tools exist that aim to facilitate the structured exchange of threat information, including, e.g., the Open Indicators of Compromise (OpenIOC) framework, Structured Threat Information Expression (STIX), Trusted Automated eXchange of Indicator Information (TAXII), the Open Vulnerability Assessment Language (OVAL), Malware Attribute Enumeration and Characterization (MAEC), the Vocabulary for Event Recording and Incident Sharing (VERIS), MISP, and MITRE’s ATT&CK framework for adversarial tactics and techniques.

Research problem:

The goal of this thesis is to review the state-of-the-art in threat intelligence sharing platforms, tools and practice. To this end, you will organize the landscape of existing tools, standards and platforms, compare and contrast their coverage and functionality, and identify strengths, weaknesses, overlaps and opportunities for integration.

Initial references:

1. W. Tounsi, H. Rais, “A survey on technical threat intelligence in the age of sophisticated cyber attacks”, Computers & Security, vol. 72, pp. 212-233, 2018. Doi: https://doi.org/10.1016/j.cose.2017.09.001.

2. Wagner, Cynthia, et al. "Misp: The design and implementation of a collaborative threat intelligence sharing platform." Proceedings of the 2016 ACM on Workshop on Information Sharing and Collaborative Security. 2016.

3. V. Mavroeidis and S. Bromander, "Cyber Threat Intelligence Model: An Evaluation of Taxonomies, Sharing Standards, and Ontologies within Cyber Threat Intelligence," 2017 European Intelligence and Security Informatics Conference (EISIC), Athens, 2017, pp. 91-98, doi: 10.1109/EISIC.2017.20

 

Threat intelligence sharing standards, e.g., OpenIOC, STIX, TAXII, OVAL, MAEC, VERIS, ATT&CK, MISP etc.

Type:                           Bachelor

Background: Benchmark datasets are the basis for a fair comparison and validation of computations. There are multiple ways to generate a benchmark dataset, including the use of existing real-world data and the generation of synthetic data. Typically, we would want benchmark datasets of various sizes (small to larger), reflecting different problem domains, and representing different levels of complexity. In this thesis, the focus will be on graph data, particularly synthetically generated RDF data sets. In this domain, synthetic benchmarks such as WatDiv [1] have become widely used in the community.

Research problem:

Multiple approaches to generate artificial data exist, it is not always clear how well they represent the real world. In the context of knowledge graphs, the extent to which synthetically generated RDF graph data sets are representative of real-world knowledge graph instances has been challenged in the literature early on [3]. More recently, analyses to characterize RDF data sets from a graph analytic perspective have used metrics from network science and applied them to knowledge graphs [6]. In this project, you will (i) collect and select graph measures and metrics [2] that can be applied to RDF graphs, (ii) collect synthetic datasets or generate them with available tools, and iii) compare the collected graphs and measure how realistic they are according to your metrics.

Initial references:

1. G. Aluç, O. Hartig, M. T. Özsu and K. Daudjee. Diversified Stress Testing of RDF Data Management Systems. In Proc. The Semantic Web - ISWC 2014 - 13th International Semantic Web Conference, 2014, pages 197-212. WatDiv available from http://dsg.uwaterloo.ca/watdiv/

2. Wills, Peter, and François G. Meyer. "Metrics for graph comparison: a practitioner’s guide." PloS one 15.2 (2020): e0228728. Available: https://arxiv.org/abs/1904.07414

3. Duan, Songyun, et al. "Apples and oranges: a comparison of RDF benchmarks and real RDF datasets." Proceedings of the 2011 ACM SIGMOD International Conference on Management of data. 2011.

4. Angela Bonifati, Irena Holubová, Arnau Prat-Pérez, and Sherif Sakr. 2020. Graph Generators: State of the Art and Open Challenges. ACM Comput. Surv. 53, 2, Article 36 (April 2020). https://doi.org/10.1145/3379445  

5. A Software Framework and Datasets for the Analysis of Graph Measures on RDF Graphs, Matthäus Zloch, Maribel Acosta, Daniel Hienert, Stefan Dietze, Stefan Conrad. 2019. https://arxiv.org/pdf/1907.01885.pdf

6. Fernández, Javier D., et al. "Characterizing RDF datasets." Journal of Information Science 1 (2016): 1-27. https://ai.wu.ac.at/qadlod/docs/JIS-3757-accepted.pdf

 

Background: 

Semantic technologies and knowledge graphs have seen widespread adoption in the web industry, where they are increasingly used in applications such as contextual search, user profiling, social network modeling, and question-answering systems (e.g., virtual voice assistants). Industry uptake in the manufacturing domain has been less pronounced, but has been gaining momentum recently as indicated by a growing number of workshops and contributions on the topic. Hence, the potential of knowledge graphs to tackle the many and varied data integration, interoperability, and AI challenges that the manufacturing domain faces in the context of Industry 4.0 is increasingly being recognized.

Research problem:

Standardized schemas, conceptualizations, and vocabularies are a key building block in the process of creating (internal or public) industrial knowledge graphs, as they facilitate integration of data from various sources within the production environment. The goal of this thesis is to review the state-of-the art in semantic representation in Industry 4.0. To this end, you will conduct research into common categories of manufacturing data and (i) identify candidate standards, schemas, and vocabularies to represent data within these categories, (ii) discuss their potential to facilitate exchange of manufacturing data and/or to break down manufacturing data silos, and (iii) identify important gaps in available vocabularies necessary to build manufacturing knowledge graphs. You will thereby organize the landscape of existing standards, compare and contrast their coverage, identify overlaps and gaps, and analyze strengths and weaknesses.

Initial references:

1. Hogan, A., Blomqvist, E., Cochez, M., D’Amato, C., Melo, G.D., Gutiérrez, C., Gayo, J.E., Kirrane, S., Neumaier, S., Polleres, A., Navigli, R., Ngomo, A.N., Rashid, S.M., Rula, A., Schmelzeisen, L., Sequeda, J., Staab, S., & Zimmermann, A. (2020). Knowledge Graphs. ArXiv, abs/2003.02320. https://arxiv.org/abs/2003.02320

2. Varish Mulwad and Raghava Mutharaju. 2017. Industrial Knowledge Graphs Workshop 2017: co-located with the 9th International ACM Web Science Conference 2017 (Preface; part of Content). In Proceedings of the 2017 ACM on Web Science Conference (WebSci '17). ACM, New York, NY, USA, 423-424. DOI: https://doi.org/10.1145/3091478.3162382

3. I. Grangel-González, L. Halilaj, S. Auer, S. Lohmann, C. Lange and D. Collarana, "An RDF-based approach for implementing industry 4.0 components with Administration Shells," 2016 IEEE 21st International Conference on Emerging Technologies and Factory Automation (ETFA), Berlin, 2016, pp. 1-8, doi: https://doi.org/10.1109/ETFA.2016.7733503.

4. P. Adolphs, H. Bedenbender, D. Dirzus, M. Ehlich, U. Epple, M. Hankel, et al., "Reference architecture model industrie 4.0 (rami4.0)", ZVEI and VDI Status Report, 2015. Available: https://www.zvei.org/fileadmin/user_upload/Presse_und_Medien/Publikationen/2016/januar/GMA_Status_Report__Reference_Archtitecture_Model_Industrie_4.0__RAMI_4.0_/GMA-Status-Report-RAMI-40-July-2015.pdf

 

Additional example references:

• N. Petersen, L. Halilaj, I. Grangel-González, S. Lohmann, C. Lange, and S. Auer. ”Realizing an RDF-based information model for a manufacturing company – A case study”. In: The Semantic Web – ISWC 2017, pages 350–366. Springer International Publishing, 2017. Available: https://link.springer.com/chapter/10.1007/978-3-319-68204-4_31

• Mörzinger, Benjamin, et al. "A large-scale framework for storage, access and analysis of time series data in the manufacturing domain." Procedia CIRP 67 (2018): 595-600.
  Available: https://www.sciencedirect.com/science/article/pii/S2212827117312131

• Annane, Amina, Nathalie Aussenac-Gilles, and Mouna Kamel. "BBO: BPMN 2.0 Based Ontology for Business Process Representation." 20th European Conference on Knowledge Management (ECKM 2019). 2019.

• Lemaignan, Severin, et al. "MASON: A proposal for an ontology of the manufacturing domain." IEEE Workshop on Distributed Intelligent Systems: Collective Intelligence and Its Applications (DIS'06). IEEE, 2006.
  Available: https://academia.skadge.org/publis/Lemaignan2006.pdf [early work]

• Lin, Hsiao-Kang, and Jenny A. Harding. "A manufacturing system engineering ontology model on the semantic web for inter-enterprise collaboration." Computers in Industry 58.5 (2007): 428-437.
  Available: https://www.sciencedirect.com/science/article/pii/S0166361506001837  [early work]

Writing smart contract code is difficult due to the various specifics of blockchain technology. For this reason, several approaches have been proposed to generate smart contract code from more user-friendly languages.

The research problem of this thesis is to implement a transformation from the Factscript language to Ethereum. To this end, the various constructs of the Factscript have to be studied and mapped to corresponding Ethereum Solidity primitives.

References

Weber, I., Xu, X., Riveret, R., Governatori, G., Ponomarev, A., & Mendling, J. (2016, September). Untrusted business process monitoring and execution using blockchain. In International Conference on Business Process Management (pp. 329-347). Springer, Cham.

Maxim Vidgof, Philipp Waibel, Jan Mendling, Martin Schimak, Alexander Seik and Peter Queteschiner: A Process Scripting and Execution Environment, ER Proceedings 2020.

López‐Pintado, O., García‐Bañuelos, L., Dumas, M., Weber, I., & Ponomarev, A. (2019). Caterpillar: a business process execution engine on the Ethereum blockchain. Software: Practice and Experience, 49(7), 1162-1193.

Xu, X., Weber, I., & Staples, M. (2019). Architecture for blockchain applications (pp. 1-307). Heidelberg: Springer.

Blockchain can be a component of a larger software system (Xu et al., 2019). Oracles are needed
for communication between the blockchain and the off-chain environment, including the other
components of the system. (Mühlberger et al., 2020) However, these oracles assume that the offchain
components are automated and provide well-defined APIs, which is not always the case.
Indeed, many companies still have legacy systems as part of their technology stack.
A possible way to automate interaction with the legacy systems is Robotic Process Automation.
Robotic Process Automation (RPA) is an emerging technology that employs software robots to
mimic human actions. (Syed et al., 2020) RPA, although essentially being a program, can interact
with the legacy systems as if it was a user, which makes it a promising solution to bridge the
communication gap between different components of a software system.

In this paper, the student will implement an oracle that would enable the communication between
the blockchain and a legacy system with the help of RPA. Note that programming skills are a strict
requirement.

For further questions, please contact Maxim Vidgof ( maxim.vidgof@wu.ac.at ).

Starting literature:
Xu, X., Weber, I., & Staples, M. (2019). Architecture for blockchain applications. Heidelberg:
Springer.
Mühlberger, R., Bachhofner, S., Ferrer, E. C., Di Ciccio, C., Weber, I., Wöhrer, M., & Zdun, U.
(2020, September). Foundational Oracle Patterns: Connecting Blockchain to the Off-chain World.
In International Conference on Business Process Management (pp. 35-51). Springer, Cham.
Syed, R., Suriadi, S., Adams, M., Bandara, W., Leemans, S. J., Ouyang, C., ter Hofstede, A. H. M.,
van de Weerd, I., Wynn, M. T. & Reijers, H. A. (2020). Robotic Process Automation: Contemporary
themes and challenges. Computers in Industry, 115, 103162.

Background:

The quality of the outcomes of process mining techniques strongly relies on the adequate input data. Domain knowledge plays an important role on the task of choosing and preparing the data. Usually process mining techniques rely on domain expert knowledge which can be error prone and not always available. Knowledge representation artifacts such as Ontology or Knowledge Graphs may support the process mining techniques

Research problem:

The core research problem addressed is: How process mining techniques can benefit from domain knowledge artifacts, such as ontologies or knowledge graphs?

This thesis can be conducted in two ways:

• With the aim of doing a systematic review of the literature on techniques that consider domain knowledge artifacts to support process mining.

• With the aim of proposing a method that uses knowledge graph for supporting process mining

Requirements:

The candidate must have previous knowledge on process mining and knowledge graph. Further desirable requirements are pro-activity and self-organization. 

Initial references

1. Mohammad Khanbabaei, Farzad Movahedi Sobhani, Mahmood Alborzi, Reza Radfar: Developing an integrated framework for using data mining techniques and ontology concepts for process improvement. J. Syst. Softw. 137: 78-95 (2018).

2. Stefano Bistarelli, Tommaso Di Noia, Marina Mongiello, Francesco Nocera:
   PrOnto: an Ontology Driven Business Process Mining Tool. KES 2017: 306-315

 

Background:

In Industry 4.0, the use of smart technologies and sensors enhances the productivity of manufacturing technologies. Process mining technologies can be used in the context of the smart manufacturing process as data analysis tool providing insights about the process to the analysts.

Research problem:

The core research problem addressed is: How process mining can be applied to the smart manufacturing industry?

This thesis can be conducted in two ways:

• With the aim of doing a systematic review of the literature on methods that apply process mining to the smart manufacturing industry..

• With the aim of proposing a method to apply process mining techniques to smart manufacturing industry

Requirements:

The candidate must have previous knowledge on process mining. Further desirable requirements are pro-activity and self-organization. 

Initial references

1. Cristina-Claudia Osman, Ana-Maria Ghiran: When Industry 4.0 meets Process Mining. KES 2019: 2130-2136

2. Mannhardt, F., Petersen, S. & Oliveira, M. Process Mining and Privacy in Smart Manufacturing. Informatik Spektrum 42, 336–339 (2019). doi.org/10.1007/s00287-019-01199-6
    

 

Background:

Process monitoring is one of the phases in the Business Process Management (BPM) life cycle. One possible task in this phase is process prediction, which predicts process indicators, such as the remaining time for the process instance to finish or the next event to occur. Prediction allows the anticipation of possible problems and they are usually based on activity information. However, in many situations the reason for the prediction is not clear to the analyst. Without explanation of the predicted outcome, the analyst does not have the necessary support for decision making.

Research problem:

The core research problem addressed is: how explainable AI can be used in the context of business process prediction?

This thesis focuses on a literature review aimed at investigating how explainable AI can be applied to business process prediction.

Requirements:

The candidate must have previous knowledge on process mining. Further desirable requirements are pro-activity and self-organization.  

Initial references

1. Marquez-Chamorro, A. E., Resinas, M., & Ruiz-Cortes, A. (n.d.). Predictive monitoring of business processes: a survey. IEEE Transactions on Services Computing, 1. doi.org/10.1109/TSC.2017.2772256

2. Alejandro Barredo Arrieta, Natalia Díaz Rodríguez, Javier Del Ser, Adrien Bennetot, Siham Tabik, Alberto Barbado, Salvador García, Sergio Gil-Lopez, Daniel Molina, Richard Benjamins, Raja Chatila, Francisco Herrera: Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI. Inf. Fusion 58: 82-115 (2020)

 

Visual analytics is a research area that has developed various techniques to analyze complex data. A particular area of visual analytics focuses on event sequence data, data that is also used in process mining where it is called event logs. Current approaches of process mining make extensive use of so-called directly-follows graph representation (also known as heuristic nets). The benefit of such a graph is that it shows all directly-follows relationships that are observed in the event log. This representation becomes less useful in the situation when the relations have uncertainty associated with them.

The objective of this thesis is to define and implement a refined visualization for directly-follows graphs in the case of additional uncertainty information. The strengths and weaknesses of a new design should be discussed based on various examples that are visualized using the prototypical implementation.

References:

Yi Guo, Shunan Guo, Zhuochen Jin, Smiti Kaul, David Gotz, Nan Cao: Survey on Visual Analysis of Event Sequence Data, https://arxiv.org/abs/2006.14291.

Sander J. J. Leemans, Erik Poppe, and Moe Thandar Wynn. 2019. Directly Follows-Based Process Mining: Exploration & a Case Study. In ICPM. IEEE, 25–32.

Yeshchenko, A., Di Ciccio, C., Mendling, J., & Polyvyanyy, A. (2019, November). Comprehensive process drift detection with visual analytics. In International Conference on Conceptual Modeling (pp. 119-135). Springer, Cham.

Satellite images offer a novel source of data about economic activity and development in places and regions. Data since the early 90s (specifically 1992-2018) is now available, and comes in a variety of formats and aggregations. The intensity of night-time lights, in particular, has been used to study economic development and activity. To date, this data is difficult to handle and to aggregate.

The primary goal of the thesis will be to create a data pipeline that ingests:
1. Longitudinal night time lights satellite data
2. EU national and subnational (NUTS 1,2,3, & municipalities) geographical borders
and outputs time series of night time light intensity statistics within these borders. The resulting outputs will be shared in an appropriate open data portal.

The following research questions can be investigated to examine the potential of this data to predict and explain economic data.
1. Does a change in light intensity correlate with economic growth?
2. Are lights a leading or lagging indicator of GDP?
3. Can lights identify economic shocks?

References:
Data source: Li, X., Zhou, Y., Zhao, M. et al. A harmonized global nighttime light dataset 1992–2018. Sci Data 7, 168 (2020). https://doi.org/10.1038/s41597-020-0510-y
Overview: https://blogs.worldbank.org/sustainablecities/nighttime-lights-are-revolutionizing-way-we-understand-covid-19-and-our-world
Archila Bustos, M.F., Hall, O. & Andersson, M. Nighttime lights and population changes in Europe 1992–2012. Ambio 44, 653–665 (2015). https://doi.org/10.1007/s13280-015-0646-8
Bruederle, A. & Hodler. R. Nighttime lights as a proxy for human development at the local level. PLOS ONE (2018).

Researchers are increasingly interested in understanding the demographics of participants in online communities, for instance GitHub, Twitter, or Stack Overflow. One key demographic of users is their location. Researchers often turn to geocoding APIs to transform self-reported location strings from user profiles in geographic data. However, these APIs (OSM, Bing Maps, Google Maps, etc.) are often optimized for traffic and wayfinding applications. As a result, researchers often apply the most convenient APIs when geolocating users from online communities, without really knowing if it is well-suited for their use case.

The goal of this thesis will be to survey geocoding solutions and APIs and to evaluate their accuracy on data sourced from online communities. This will require a significant manual geocoding effort to generate a gold corpus of ground truth locations.

 Additional areas to investigate include:
- Classifying clearly fake locations using machine learning (i.e. "The Moon", "in the Universe").
- Classifying double locations using machine learning ("I split my time between Tokyo and New York").
- Evaluating the performance of individual APIs vs an ensemble geocoding effort: can data from multiple APIs significantly outperform the best single API?
- Setting up a Nominatim (OpenStreetMap API) server at WU for local use.

References:
Mourad, A., Scholer, F., Magdy, W., & Sanderson, M. (2019). A practical guide for the effective evaluation of twitter user geolocation. ACM Transactions on Social Computing, 2(3), 1-23.
Hecht, B., Hong, L., Suh, B., & Chi, E. H. (2011, May). Tweets from Justin Bieber's heart: the dynamics of the location field in user profiles. In Proceedings of the SIGCHI conference on human factors in computing systems (pp. 237-246).
Kariryaa, A., Johnson, I., Schöning, J., & Hecht, B. (2018, April). Defining and predicting the localness of volunteered geographic information using ground truth data. In Proceedings of the 2018 CHI conference on human factors in computing systems (pp. 1-12).
Takhteyev, Y., & Hilts, A. (2010). Investigating the geography of open source software through GitHub. Manuscript submitted for publication.

Coding skills required

Holacracy (Robertson, 2015; Bernstein, Bunch, Lee and Canner, 2016) is a form of self-management that relies on IT-based formalization to enable decentralized decision making. While there are many claims surrounding this New Form of Organizing (Puranam, Alexy and Reitzig, 2014), there are little empirical insights on how holacratic organizing really works.

In this thesis, you should investigate the power network of a holacrtic organization. We have digital trace data (Berente, Seidel and Safadi, 2019) that contains information on how a holacratic organization developed over the course of seven years. You should use dynamic network and social network analysis (Moody, McFarland and Bender‐deMoll, 2005) to investigate how the organizational structure changed during this period.

For further questions, please contact Bastian Wurm (bastian.wurm@wu.ac.at).

Starting Literature:

Berente, N., S. Seidel and H. Safadi. (2019). “Data-Driven Computationally-Intensive Theory Development.” Information Systems Research, 30(1), iii–viii.

Bernstein, E., J. Bunch, M. Lee and N. Canner. (2016). “Beyond the Holacracy Hype.” Harvard Business Review, 94(7), 38–49.

Moody, J., D. McFarland and S. Bender‐deMoll. (2005). “Dynamic Network Visualization.” American Journal of Sociology, 110(4), 1206–1241.

Puranam, P., O. Alexy and M. Reitzig. (2014). “What’s “new” about new Forms of Organizing?” Academy of Management Review, 39(2), 162–180.

Robertson, B. J. (2015). Holacracy: The new management system for a rapidly changing world. Henry Holt and Company.

Nowadays, business processes are an integral part of almost every organization. The companies are thereby supported by information systems that provide a means to manage and monitor the execution of the processes. To store and query process relevant data, these information systems are often relying on relational databases. However, in comparison to relational databases, newer graph-based databases might provide more native support to store process relevant data. This stems from the fact that graph-based databases natively support the storage and querying of graphs, which processes, in their simplest form, are.

The research problem of this thesis is to investigate how and to which extend graph-based databases can be used to store business process relevant data. For this, the literature has to be analyzed in the field of business processes and graph-based databases. Building on this literature research, an evaluation of different graph-based databases for the suitability of storing and querying of business process relevant data has to be done.

References

Esser, S., & Fahland, D. (2020). Multi-Dimensional Event Data in Graph Databases. arXiv preprint arXiv:2005.14552.

Waibel, P., Novak, C. Bala, S., Revoredo, K., and Mendling, J.. Analysis of Business Process Batching using Causal Event Models. In EdbA’20 workshop at ICPM 2020.

Hogan, A., Blomqvist, E., Cochez, M., d'Amato, C., de Melo, G., Gutierrez, C., et. al & Zimmermann, A. (2020). Knowledge graphs. arXiv preprint arXiv:2003.02320.

Kumar Kaliyar, R. (2015, May). Graph databases: A survey. In International Conference on Computing, Communication & Automation (pp. 785-790). IEEE.

Background:

Solving problems often requires seeing new connections between concepts or events that seemed unrelated at first. (Adam E. Green, 2010) Innovative solutions of this kind depend on analogical reasoning, a relational reasoning process that involves mapping similarities between concepts. Brain-based evidence has implicated the frontal pole of the brain as important for analogical mapping. Separately, cognitive research has identified semantic distance as a key characteristic of the kind of analogical mapping that can support innovation (i.e., identifying similarities across greater semantic distance reveals connections that support more innovative solutions and models). More recently, the analogical reasoning has attracted a lot of attention in computer sciences, especially artificial intelligence. (Henri Prade, 2014)

Research problem:

The goal of this thesis is to (1) review and (2) organize the state-of-the-art in analogical reasoning in computer sciences with the main guiding question of: How is analogical reasoning being digitally implemented?

Initial references:

1. Green, A. E., Kraemer, D. J., Fugelsang, J. A., Gray, J. R., & Dunbar, K. N. (2010). Connecting long distance: semantic distance in analogical reasoning modulates frontopolar cortex activity. Cerebral cortex, 20(1), 70-76.

 

https://academic.oup.com/cercor/article/20/1/70/413473?login=true

1. Prade, H., & Richard, G. (Eds.). (2014). Computational approaches to analogical reasoning: Current trends (Vol. 548). Heidelberg: Springer.

 

https://link.springer.com/content/pdf/10.1007/978-3-642-54516-0.pdf

1. Helman, D. H. (Ed.). (2013). Analogical Reasoning: perspectives of artificial intelligence, cognitive science, and philosophy (Vol. 197). Springer Science & Business Media.

 

https://books.google.at/books?hl=ko&lr=&id=Ue3qCAAAQBAJ&oi=fnd&pg=PA1&dq=analogical+reasoning+computer&ots=Wsi6O8913C&sig=9i2sQAbfwkJLViGmzhvAv5mR7Tg#v=onepage&q=analogical%20reasoning%20computer&f=false

Recent years have seen an increasing availability of process execution data from several data sources. Process mining offers different analysis techniques to extract business insights from these data, known as event logs. An event log is a set of the executed process instances, i.e., cases. Each event in the event log represents an execution of process activity. The primary attributes associated with an event carry the control-flow information as execution timestamp, executed activity, and case identifier. Additionally, an event has different data attributes that describe business objects used for the process execution and process environment, e.g., resources, project name, location, cost..., etc.

The analysts use the process mining techniques to discover the control-flow model, identify the deviations using conformance checking techniques and provide recommendations and suggestions for process improvement.

Unfortunately, these techniques focus on the event process-attributes and ignoring the event data-context attributes. Analyzing the data-context attributes will improve the understanding of the process execution environment and the performed analysis quality.

The data-context attributes provide information about the business objects affected by executing the process instances. These attributes may be constant within a case or change over time in a case. Detecting the data-context attributes changing patterns will help understand what triggers these changes and how they affect the process variations. To detect the changing patterns, we need to identify the association and correlation rules between the data-context attributes. 

The thesis's objective is to build a tool that detects the associate and correlation rules between the log's event data-context attributes.

References:

[1] Pentland, B., et al. "Bringing context inside process research with digital trace data." Journal of the Association for Information Systems 21 (2020).

[2] Pegoraro, Marco, Merih Seran Uysal, and Wil MP van der Aalst. "Discovering Process Models from Uncertain Event Data." International Conference on Business Process Management. Springer, Cham, 2019.

[3] Motahari-Nezhad, Hamid Reza, et al. "Event correlation for process discovery from web service interaction logs." The VLDB Journal 20.3 (2011): 417-444.

[4] Hornik, Kurt, Bettina Grün, and Michael Hahsler. "arules-A computational environment for mining association rules and frequent item sets." Journal of Statistical Software 14.15 (2005): 1-25.

Managing and preserving knowledge in the best possible way has always been a key to success for organizations, even long before the term “Knowledge Graph” entered the stage. However, with the rise of practical applications (such as Google´s announcement of their knowledge graph in 2012 [1]) and the growing body of literature on knowledge graphs [2], one might wonder about the connection between the term “knowledge” from a Knowledge Management perspective and the knowledge actually represented in a knowledge graph (from the viewpoint of the complementary research field of Knowledge Representation and Reasoning). Inevitably this leads to the question, how much “knowledge” there actually is in a Knowledge Graph! In this thesis your task is to explore the answer to this question by conducting a systematic literature survey on publications that describe the application of KGs in specific industry domains, where you should explore the academic literature, but also announcements of the usage of Knowledge Graphs in industry [3] in finding evidence in . You should consider articles from scientific journals as well as conference proceedings and try to identify those papers that deal with knowledge graphs and knowledge creation, with particular focus on the exchange between implicit/tacit and explicit knowledge [4]. Besides the identification of articles about this subject you can also consider, how knowledge graphs can contribute to knowledge management and vice versa.

Your work will be part of an interdisciplinary research project on the subject which also includes the chance for you to publish parts of your thesis, or be part of an extended version of a joint paper we are currently working on [5].

Supervisor(s):

Clemens Kerschbaum, Axel Polleres, Rositsa Ivanova, Alexander Kaiser

Literature:   

1. https://blog.google/products/search/introducing-knowledge-graph-things-not/ , blogpost published 2012

2. Aidan Hogan, Eva Blomqvist, Michael Cochez, Claudia d'Amato, Gerard de Melo, Claudio Gutierrez, José Emilio Labra Gayo, Sabrina Kirrane, Sebastian Neumaier, Axel Polleres, Roberto Navigli, Axel-Cyrille Ngonga Ngomo, Sabbir M. Rashid, Anisa Rula, Lukas Schmelzeisen, Juan Sequeda, Steffen Staab, and Antoine Zimmermann. Knowledge graphs. ACM Computing Surveys (CSUR), accepted for publication, 2021. to appear. Extended pre-print available at https://arxiv.org/abs/2003.02320.

3. For a collection of company names that use Knowledge Graphs, see for instance: https://www.slideshare.net/Frank.van.Harmelen/adoption-of-knowledge-graphs-late-2019  

4. I. Nonaka, R. Toyama and N. Konno, SECI, Ba and leadership: a unified model of dynamic knowledge creation, Long range planning 33(1) (2000), 5–34.

5. R.V. Ivanova, C. Kerschbaum, S. Neumaier, A. Kaiser, A. Polleres, How much Knowledge is in Knowledge Graphs? - A Knowledge Management Perspective, unpublished manuscript.

The tracking of issues in software systems by using so-called Issue Tracking Systems (ITS) is an approach to record issues and to follow the progress of resolving them. While this approach is widely used in software engineering, the usage of ITS in the field of Process Mining and Process Analytics, e.g., to track issues in processes and the resolving of them, were so far neglected.

The research problem of this thesis is to investigate which techniques and approaches of ITSs are particularly suitable for process mining. To this end, the literature on the usage of ITS has to be analyzed to identify potential use cases for process mining. To further investigate which value it could bring to a process analyst, the result of this literature research should then be evaluated by showing how it could be applied in combination with existing process mining tools.

References

Bertram, D., Voida, A., Greenberg, S., & Walker, R. (2010). Communication, collaboration, and bugs: the social nature of issue tracking in small, collocated teams. In Proceedings of the 2010 ACM conference on Computer supported cooperative work (pp. 291-300).

Lessel, P., Müller, M., & Krüger, A. (2015). Towards a novel issue tracking system for" industry 4.0" environments. In Proceedings of the 33rd Annual ACM Conference Extended Abstracts on Human Factors in Computing Systems (pp. 1809-1814).

Van Der Aalst, W., Adriansyah, A., De Medeiros, A. K. A., Arcieri, F., Baier, T., Blickle, T., et al. & Wynn, M. (2011). Process mining manifesto. In International Conference on Business Process Management (pp. 169-194).

Process mining is devoted to the extraction of value from process data. An important research stream is process discovery: building a process model from the execution traces (event logs) only. The majority of efforts in this direction have been aimed at the discovery of procedural models, either in formal representations, e.g., Petri-nets, or higher level notations, e.g., BPMN.

Diagonal to the procedural paradigm is the declarative one: only the boundaries of the process are defined, not the specific control flow, allowing any action which does not contradict these constraints. This is a flexible representation which is gaining more and more attention, thanks to its power to facilitate the handling of complex and dynamic processes where the order of actions changes in each execution.

Following this demand, the Object Management Group (OMG) released the Case Management Model and Notation (CMMN), an industry standard notation for declarative specifications.

Research problem:

Despite being a standard since 2014, the research on CMMN is still scarce while the efforts are focused on notations like DECLARE or DCR graphs. These are well researched and known in the academic community, where they originated, but struggle to break into industry.

The thesis goal is to fill this gap developing a technique for automated discovery of CMMN models from event log data and its proof-of-concept software implementation.

Requirements:

Good programming skills

Initial references:

1- CMMN official specification: https://www.omg.org/cmmn/

2- Survey on Declarative process mining techniques:

Tijs Slaats: Declarative and Hybrid Process Discovery: Recent Advances and Open Challenges. J. Data Semant. 9(1): 3-20 (2020)

3 - Survey on automated process discovery techniques (procedural and declarative):

Adriano Augusto, Raffaele Conforti, Marlon Dumas, Marcello La Rosa, Fabrizio Maria Maggi, Andrea Marrella, Massimo Mecella, Allar Soo: Automated Discovery of Process Models from Event Logs: Review and Benchmark. IEEE Trans. Knowl. Data Eng. 31(4): 686-705 (2019).

If you ever wondered how it could be possible to depict or even measure structural differences of organizations, this thesis topic could be interesting for you. In this thesis, your task is to do a literature review on ways to describe organizations, with special emphasize on their structure (i.e., the way they work). You should identify and summarize typical parameters or characteristics that can be used to describe the structure of how work is done, how learning takes place or how the organization decides on its strategy for the future. If you are interested in this topic, please be aware that you will find yourself confronted with a great amount of existing literature on this subject. Therefore, a big part of your work will be to select the most relevant from a pool of many other publications.

Literature:

Birnholtz, J. P., Cohen, M. D., & Hoch, S. V. (2007). Organizational character: On the regeneration of camp poplar grove. Organization Science, 18(2), 315-332.

Shafritz, J. M., Ott, J. S., & Jang, Y. S. (2015). Classics of organization theory. Cengage Learning.

Mintzberg, H. (1989). The structuring of organizations. In Readings in strategic management (pp. 322-352). Palgrave, London.

“We live in a VUCA world” is a sentence you are likely to find when going through recent literature on strategic management and related fields. With these words, writers usually acknowledge that the environment we live in is neither linear nor predictable. Arguably, we need a certain type of leadership, strategy or organizational structure in order to deal with those conditions. In this thesis your task will be to conduct a literature review on those certain organizational structures (i.e. organizational forms). Within this thesis you should investigate: 1) at which point in time, academic literature recognized the rise of alternative organizational forms, 2) collect different types of organizational forms mentioned in academic literature and 3) try to elaborate a structural comparison of the organizational forms you found. (VUCA = volatility, uncertainty, complexity, ambiguity)

Literature:

Lee, M. Y., & Edmondson, A. C. (2017). Self-managing organizations: Exploring the limits of less-hierarchical organizing. Research in organizational behavior, 37, 35-58.

Martela, F. (2019). What makes self-managing organizations novel? Comparing how Weberian bureaucracy, Mintzberg’s adhocracy, and self-organizing solve six fundamental problems of organizing. Journal of Organization Design, 8(1), 1-23.

Mitleton-Kelly, E. (2003). Ten principles of complexity and enabling infrastructures. Complex systems and evolutionary perspectives on organisations: The application of complexity theory to organisations, 1, 23-50.

Knowledge graphs are means to model a domain of interest via relationships between objects, where the objects are the nodes and the relationships are the edges of a graph. This is done with the intention to be knowledge (Hogan, 2020), e.g., in medicine (Rotmensch, 2017). These nodes and edges can be further mapped into a continuous vector space – which is referred to as embedding a knowledge graph (Wang, 2017). Embeddings are, however, not limited to nodes and edges but are also done for substructures (subset of nodes and/or edges) and even the whole-graph (Cai, 2018). Manufacturing companies hence might find the knowledge graph properties, i.e., explicit modelling of relationships and ability to embed the graph, helpful in achieving business goals, e.g., when multiple sensors capture data on an object which implies dependent measurements (Garofalo, 2018). 

The objective of this thesis is to conduct a comprehensive literature review on applications of knowledge graph embeddings in the context of manufacturing and Industry 4.0. In the thesis, the student will systematically identify and characterize use cases upon multiple dimensions (e.g., industry type, type of process, etc.) which might reveal interesting patterns, and even new use cases. The student might also be asked to give her/his opinion on the current state of the art. The review shall include grey literature, i.e., non-scientific pieces of literature, e.g., company websites, marketing artefacts, and the likes. 

Supervisors: Stefan Bachhofner, Elmar Kiesling 

Resources 

CS 520, Knowledge Graphs, Stanford University, https://web.stanford.edu/class/cs520/, web.stanford.edu/~vinayc/kg/notes/Table_Of_Contents.html 

References 

Hogan, A., Blomqvist, E., Cochez, M., d'Amato, C., de Melo, G., Gutierrez, C., ... & Zimmermann, A. (2020). Knowledge graphs. arXiv preprint arXiv:2003.02320. 

Rotmensch, M., Halpern, Y., Tlimat, A., Horng, S., & Sontag, D. (2017). Learning a health knowledge graph from electronic medical records. Scientific reports, 7(1), 1-11. 

Wang, Q., Mao, Z., Wang, B., & Guo, L. (2017). Knowledge graph embedding: A survey of approaches and applications. IEEE Transactions on Knowledge and Data Engineering, 29(12), 2724-2743. 

Cai, H., Zheng, V. W., & Chang, K. C. C. (2018). A comprehensive survey of graph embedding: Problems, techniques, and applications. IEEE Transactions on Knowledge and Data Engineering, 30(9), 1616-1637. 

Garofalo, M., Pellegrino, M. A., Altabba, A., & Cochez, M. (2018). Leveraging knowledge graph embedding techniques for industry 4.0 use cases. arXiv preprint arXiv:1808.00434. 

Deep Learning has had tremendous success in business and science (LeCun, 2015, Schmidhuber, 2015, Bachhofner, 2020). The paper by (Krizhevsky, 2012) is widely accepted to be the key paper for this success, as they were able to substantially decrease the error on ImageNet Large Scale Visual Recognition Challenge 2012 (ILSVRC-2012), an image classification task, by using a convolutional neural network. Since this work, many scholars worked on scaling neural networks based on the idea that larger models lead to a better performance. One such scaling idea is the compound coefficient proposed by (Tan, 2019), which uniformly scales depth, width, and resolution. 

In this thesis, the student contributes to the question whether the compound coefficient has an influence on error patterns. In particular, to answer this question the students task is to perform a series of reproducible image classification experiments. 

Resources 

STAT 157, Introduction to Deep Learning, UC Berkley, https://courses.d2l.ai/berkeley-stat-157/index.html 

6.S191, Introduction to Deep Learning, Massachusetts Institute of Technology, introtodeeplearning.com/ 

Dive into Deep Learning, d2l.ai/ 

CS231n, Convolutional Neural Networks for Visual Recognition, Stanford University, http://cs231n.stanford.edu/ 

A collection of lecture on deep learning, Massachusetts Institute of Technology, deeplearning.mit.edu/ 

References 

LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521(7553), 436-444. 

Schmidhuber, J. (2015). Deep learning in neural networks: An overview. Neural networks, 61, 85-117. 

Bachhofner, S., Loghin, A. M., Otepka, J., Pfeifer, N., Hornacek, M., Siposova, A., ... & Hochreiter, R. (2020). Generalized Sparse Convolutional Neural Networks for Semantic Segmentation of Point Clouds Derived from Tri-Stereo Satellite Imagery. Remote Sensing, 12(8), 1289. 

Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks. Advances in neural information processing systems, 25, 1097-1105. 

Bilal, A., Jourabloo, A., Ye, M., Liu, X., & Ren, L. (2017). Do convolutional neural networks learn class hierarchy?. IEEE transactions on visualization and computer graphics, 24(1), 152-162. 

Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., ... & Rabinovich, A. (2015). Going deeper with convolutions. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 1-9). 

Tan, M., & Le, Q. (2019, May). Efficientnet: Rethinking model scaling for convolutional neural networks. In International Conference on Machine Learning (pp. 6105-6114). PMLR.

Deep Learning has had tremendous success in business and science (LeCun, 2015, Schmidhuber, 2015, Bachhofner, 2020). The paper by (Krizhevsky, 2012) is widely accepted to be the key paper for this success, as they were able to substantially decrease the error on ImageNet Large Scale Visual Recognition Challenge 2012 (ILSVRC-2012), an image classification task, by using a convolutional neural network. One of the key ideas the authors use is to model neurons as Rectifier Linear Units (ReLU). Among other things, authors state that they observed a significant speed up due to this neuron model. (Bilal, 2017) observed that errors by GoogleLeNet (Szegedy, 2015), an Inception based architecture, to not follow a random process but suggest that it implicitly learns the class hierarchy present in the output classes (the last layer of the network). 

In this thesis, the student contributes to the question whether different models of a neuron lead to different error patterns. In particular, to answer this question the students task is to perform a series of reproducible image classification experiments. 

Resources 

STAT 157, Introduction to Deep Learning, UC Berkley, https://courses.d2l.ai/berkeley-stat-157/index.html 

6.S191, Introduction to Deep Learning, Massachusetts Institute of Technology, introtodeeplearning.com/ 

Dive into Deep Learning, d2l.ai/ 

CS231n, Convolutional Neural Networks for Visual Recognition, Stanford University, http://cs231n.stanford.edu/ 

A collection of lecture on deep learning, Massachusetts Institute of Technology, deeplearning.mit.edu/ 

References 

LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521(7553), 436-444. 

Schmidhuber, J. (2015). Deep learning in neural networks: An overview. Neural networks, 61, 85-117. 

Bachhofner, S., Loghin, A. M., Otepka, J., Pfeifer, N., Hornacek, M., Siposova, A., ... & Hochreiter, R. (2020). Generalized Sparse Convolutional Neural Networks for Semantic Segmentation of Point Clouds Derived from Tri-Stereo Satellite Imagery. Remote Sensing, 12(8), 1289. 

Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks. Advances in neural information processing systems, 25, 1097-1105. 

Bilal, A., Jourabloo, A., Ye, M., Liu, X., & Ren, L. (2017). Do convolutional neural networks learn class hierarchy?. IEEE transactions on visualization and computer graphics, 24(1), 152-162. 

Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., ... & Rabinovich, A. (2015). Going deeper with convolutions. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 1-9).

Teaching and tutoring are means for transferring knowledge. While passing  knowledge  “from a mind to another mind for the purpose of continuation of life” (Mustafa, 2018) remains unchanged, (recent) technology-induced innovations are altering learning interactions and knowledge sharing. In this, digital learning platforms are seen as key to transforming professional learning environments.

Nonaka et al. (2000) and others have emphasized the importance of the shared context in which knowledge creation takes place. More than 20 years ago, they already considered virtual spaces that can complement individual face-to-face interactions. However, when learning and knowledge creation occur entirely online, the question arises of how to transfer the vast majority of knowledge, namely tacit knowledge. This question becomes particularly important when considering professions that rely on less formalized forms of knowledge, namely téchne, which are rooted in direct experience tied to the knower.

In this bachelor thesis, you will consider different forms of knowledge and scrutinize pedagogical ways of imparting knowledge. In a structured literature review, you will provide an overview of the literature on tacit knowledge transfer in the context of online learning environments, with a particular focus on the imparting of less formalized and practical knowledge. In the second part of the thesis, you will build on the theory of ba (Nonaka et al., 2000) and propose how its theoretical propositions could be implemented in e-learning platforms, providing some examples of concrete implementation ideas (e.g., mockups, processes).

For further questions, please contact Florian Kragulj (florian.kragulj@wu.ac.at)

Initial references:

Buunk, I., Smith, C. F., & Hall, H. (2019). Tacit knowledge sharing in online environments: Locating ‘Ba’within a platform for public sector professionals. Journal of Librarianship and Information Science, 51(4), 1134-1145.

Groth, C., Mäkelä, M., & Seitamaa-Hakkarainen, P. (2015). Tactile augmentation: A multimethod for capturing experiential knowledge. Craft Research, 6(1), 57-81.

Nimkulrat, N., Niedderer, K., & Evans, M. (2015). On understanding expertise, connoisseurship, and experiential knowledge in professional practice. Journal of Research Practice, 11(2).

Mustafa, A. (2020). The personalization of e-learning systems with the contrast of strategic knowledge and learner’s learning preferences: An investigatory analysis. Applied Computing and Informatics.

Nonaka, I., Toyama, R., & Konno, N. (2000). SECI, Ba and leadership: a unified model of dynamic knowledge creation. Long range planning, 33(1), 5-34.

Tee, M. Y., & Karney, D. (2010). Sharing and cultivating tacit knowledge in an online learning environment. International Journal of Computer-Supported Collaborative Learning, 5(4), 385-413.

Wood, N., Rust, C., & Horne, G. (2009). A tacit understanding: The designer's role in capturing and passing on the skilled knowledge of master craftsmen. International Journal of Design, 3(3).

Background: Developing new methods for exploring and analyzing data in virtual and augmented reality presents many opportunities and challenges, both in terms of software development and design inspiration. There are various hardware options, from Google Cardboard to Oculus Rift. Taking part in this challenge demands programming skills as well as creativity. A basic VR or AR application for exploring a specific type of open data will be developed by the student. The use of a platform-independent kit such as A-Frame is essential, as the application will be compared in a small user study to its non-VR version in order to identify advantages and disadvantages of the visualization method implemented. Details will be discussed with supervisor. Research problem: How can AR and VR be used to improve exploration of data? Some References: Butcher, Peter WS, and Panagiotis D. Ritsos. "Building Immersive Data Visualizations for the Web." Proceedings of International Conference on Cyberworlds (CW’17), Chester, UK. 2017. Teo, Theophilus, et al. "Data fragment: Virtual reality for viewing and querying large image sets." Virtual Reality (VR), 2017 IEEE. IEEE, 2017.

Millais, Patrick, Simon L. Jones, and Ryan Kelly. "Exploring Data in
Virtual Reality: Comparisons with 2D Data Visualizations." Extended
Abstracts of the 2018 CHI Conference on Human Factors in
Computing Systems. ACM, 2018.

Organizational processes change due to various reasons (Feldman and Pentland, 2003; Dumas, La Rosa, Mendling and Reijers, 2018). This change can be approached from different perspectives. For example, Yeshchenko et al. (2019) use rule associations to visualize how organizational processes change over time.

In this thesis, you should investigate how one can approach, visualize, and measure change in organizational processes.

For further questions, please contact Bastian Wurm (bastian.wurm@wu.ac.at).

Starting Literature:

Dumas, M., M. La Rosa, J. Mendling and H. A. Reijers. (2018). Fundamentals of Business Process Management (2nd Ed.). Heidelberg: Springer.

Feldman, M. S. and B. T. Pentland. (2003). “Reconceptualizing Oragnizational Routines as a Source of Flexibility and Change.” Administrative Science Quarterly, 48(1), 94–118.

Yeshchenko, A., Di Ciccio, C., Mendling, J., & Polyvyanyy, A. (2019). Comprehensive process drift detection with visual analytics. In International Conference on Conceptual Modeling (pp. 119-135). Springer, Cham.

Recently, a lot of commonsense real-world knowledge is being accumulated as structured data in enceclopedic collections such as Wikipedia/DBpedia [1] or Wikidata [2], building up structured data graphs, also called Knowledge Graphs [3]. Many of the links captured in these knowledge graphs (such as the relations between countries and their capitals, movies and their directors, etc.), reflect relational 1:n or n:m relationships that can typically be translated to ER diagrams and - rather than RDF and graph models - basic relational schemata.

The goal of this thesis is to devise algorithms to detect and construct "instances" of certain ER model patterns within such Knowledge Graphs, in order to automatically generate small or medium-sized example schemata and instance data, for educational purposes, that can be used to auto-generate illustrative examples for database courses. You will be using tools like SPARQL and SQL to devise a conceptual solution or prototype for this purpose under our guidance.

• Auer, Sören, et al. "Dbpedia: A nucleus for a web of open data." The semantic web. Springer, Berlin, Heidelberg, 2007. 722-735.

• Vrandečić, Denny, and Markus Krötzsch. "Wikidata: a free collaborative knowledgebase." Communications of the ACM 57.10 (2014): 78-85.

• Hogan A, Blomqvist E, Cochez M, d'Amato C, de Melo G, Gutierrez C, Gayo JE, Kirrane S, Neumaier S, Polleres A, Navigli R. Knowledge graphs. arXiv preprint arXiv:2003.02320. 2020 Mar 4.

The Open Data movement, mainly driven by public administrations in the form of Open Government Data has over the last years created a rich source of structured data published on the Web, in various formats, covering dierent domains and typically available under liberal licences on so called Open Data Portals. Various studies since have attempted qualitative and quantiative assessment of Open Data initiatives and portals across different countries and portals in comparable manner [3-6]. In our own research in the institute, we have as well built up a large data collection of evolving Open data by regularly crawling data and metadata from these portals over a considerable ammount of time [1,2].

The goal of this thesis is to analyse these cralws and attempt to draw conclusions about the development of Open Data in comparison of different countries, and compare the results with other initiatives in a principled manner.

You can build up on a previous BSc thesis that has already compared different similar initiatives rather on a literature review level, whereas here the goal is to tackle the topic from a Data analytics/Data Science perspective.

1. Thomas Weber, Johann Mitlöhner, Sebastian Neumaier, and Axel Polleres. Odarchive - creating an archive for structured data from open data portals. In Proceedings of the 19th International Semantic Web Conference (ISWC 2020), volume 12507 of Lecture Notes in Computer Science (LNCS), pages 311--327, Virtual Conference (Athens, Greece), November 2020. Springer. [ .pdf ] 2.Sebastian Neumaier, Jürgen Umbrich, and Axel Polleres. Lifting data portals to the web of data. In 10th Workshop on Linked Data on the Web (LDOW2017), Perth, Austrialia, April 2017. [ .pdf ]

3. https://www.europeandataportal.eu/en/impact-studies/open-data-maturity

4. https://index.okfn.org/about/

5. https://opendatamonitor.eu/

6. https://opendatabarometer.org/

Recently, with the pandemic still enduring, and the situation becoming partially more instransparent, a lot of calls are raised for more transparent and open reporting of data around the developments of the pandemic and it's effects and development. While most countries publish obligatory dashboards and infection numbers, reporting about other factors seems to be still highly heterogeneous across countries, such that a comparable view of developments is hardly achievable.

In this, thesis the goal is to survey, overview, and compare the availability of Open Data around the pandemic across the EU countries, and assess the feasiblity of (and potentially prototype) the integration of data published by different countries in a Data Processing Pipeline ready for analytical comparisons.

Related questions to be answered include:

Which aspects of COVID related numbers on national level are publicly available in each country?

• infection numbers

• test numbers, kinds of tests used?

• R0?

• availability of testing facilities

• 7-day incidence?

• at which level of regional granularity?

• age distributions?

• vaccine numbers? Which types of vaccines? at which temporal/regional granularity? Age distribution? Risk group definitions and planned vaccine phases? Vaccine order and delivery numbers? Production numbers?

• hospitalizations and hospital capacities? ICU beds, etc.

• lockdown measures? (shops, restaurants, childcare, school closing (per grades), home office obligations)

• quarantine regulations and travel restrictions

• socio-economic effects (such as un-employment numbers, GDP effects)

How is this data published? • structured/machine-readable? • data formats • spatial and temporal granularity • available via APIs? • which schema is used

Overall goal: find out how comparable Data about covid is across EU countries or pinpoint the heterogeneity problems across the EU

The data availabilty situation assessment and integration prototype development could potentially be split up into two separate theses topics. For a prototype implementation, we are particularly interested in the deployment of Knowledge Graph technologies [1], i.e. building up a dynamic Knowledge graph on COVID-related information, that would enable or potentially enrich country- or regionwise analyses and studies such as the following examples [2,3,4], eventually.

1. https://arxiv.org/abs/2003.02320.

2. https://www.nature.com/articles/s41562-020-01009-0?fbclid=IwAR0KjYTSmexyeQnjyAsPVVriHmxVmTG5oGRgVG6oxIQe9STPLyk152lyJz4

3. https://www.nature.com/articles/s41597-020-00609-9

4. https://vis.csh.ac.at/corona-ampel/

Das folgende ist ein Literatur-Thema, bzw. ein Survey über bestehende Systeme im Bereich Forschungsdatenmanagement.

Title: Platforms for Research Data Management

Description: There is increasing pressure for more transparency, easier access and openness in scientific research, as shown by initiatives such as the Go FAIR [1] Initiative by the research data alliance (RDA), https://www.go-fair.org/, or the increasing demand for more open access to research publications and more effective management of data assets connected to research results.

In this thesis, the goal is to survey the landscape of available research data management platforms and compare them along a set of to be defined criteria. The objective is to answer the following research question "Which systems and platforms could a university like WU use to manage their research data?" by analysing and categorising the landscape of available tools and platforms in this area, find out about their adoptions, features, underlying technologies, interfaces and usability, licenses & costs, data governance principles, but also integratability in WU's current systems landscape. The latter could be also investigated by means of interviews with stakeholders at WU, whereas as a first part, research would be mostly based on online research and systmatically surveying and testing existing tools and platforms.. Available platforms and projects to establish such platforms include particularly, platforms focused on Social Sciences research (e.g. CESSDA [2], AUSSDA [3]), or platforms to pool data storing and processsing across (academic) research institutions (EOSC [4], VSC [5]) on a national or European level or many of other international similar initiatives e.g. ISPSR [6] in the US, NFDI in Germany [7] and many more... the goal of this theses is to exactly investiage the full landscape and state-of-the art in this space. Out of scope (or being subject to a complementary investigation in another thesis) would be commercial cloud offerings, that are tailored to maintain, collect and share research data. We are looking for thesis research proposals that - looking at these starting points - come up with a plan to investigate, categorize and analyze the market as well as technical solutions in the emerging sector.

1. https://www.nature.com/articles/sdata201618

2. CESSDA

3. https://aussda.at/

4. https://www.cessda.eu/

5. https://www.eosc-portal.eu/

6 https://www.icpsr.umich.edu/icpsrweb/

7 https://www.ratswd.de/forschungsdaten/fdz