Bardh Prenkaj

Bardh Prenkaj

Computer Scientist, PhD

Technical University of Munich

Biography

Currently a postdoc in ML Privacy and Security at the Technical University of Munich.

From 10/2022 to 10/2024, I was a postdoc at Sapienza in anomaly detection and explainability. During this period (June-September 2023), I spent time at my current research group, RDS working on graph counterfactual explainability.

I received my PhD in Computer Science from the Sapienza University of Rome in 2022 with a thesis on “Latent Deep Sequential Learning of Behavioural Sequences”. In Rome, I worked in the AIIM - formerly IIM - group advised by Paola Velardi and Giovanni Stilo. I then worked as a senior researcher at the Department of Computer Science in Sapienza in anomaly detection for social isolation disorders (February - September 2022).

I made BetterScholar with Antonio Norelli, an alternative to Google Scholar profiles. Read more here: Why BetterScholar?

Download my Resume.

Download my Curriculum Vitae.

Interests
  • Anomaly Detection
  • Machine Robustness
  • Machine Unlearning
  • AI for Healthcare
Education

  • PhD in Computer Science, 2022

    Sapienza University of Rome

  • MSc in Computer Science (Grade 110/110 with honors; GPA 4.00, top 1%), 2018

    Sapienza University of Rome

  • BSc in Computer Science (Grade 110/110; GPA 4.00, top 1%), 2016

    Sapienza University of Rome

Research Experience

 
 
 
 
 
Technical University of Munich
Postdoctoral Researcher
Technical University of Munich
Apr 2024 – Mar 2026 Munich, Germany
I am currently working on fairness and trustworthiness of generative models as part of the Chair of Responsible Data Science. I currently co-advise 3 PhD students.
 
 
 
 
 
Sapienza University of Rome
Postdoctoral Researcher
Sapienza University of Rome
Oct 2022 – Jan 2026 Rome, Italy

Competition procedure: AR-B 03/2022

I worked on anomaly detection in various tasks such as video understanding (in collaboration with PINlab), and behavioral and health time series (in collaboration with prof. Velardi). I also worked in counterfactual explainability in graph classification tasks (in collaboration with AIIM) where I mentored Mario Alfonso Prado-Romero (Gran Sasso Science Institute) and highly contributed to the technological transfer of GRETEL.

 
 
 
 
 
Sapienza University of Rome
Senior Research Fellow
Sapienza University of Rome
Dec 2021 – Sep 2022 Rome, Italy

Competition procedure: BS-S 6/2021

Coordinated research and implementation of innovative deep learning algorithms to predict anomalous events in patient behavioural time series.

 
 
 
 
 
George Mason University
Junior Researcher
George Mason University
Mar 2020 – Dec 2021 Faifax, VA, USA (remote)
Worked alongside Dr Sarvari and Prof Domeniconi on boosting-based anomaly detection models. We published Boosting-based Autoencoder at PAKDD'21.
 
 
 
 
 
Sapienza University of Rome
Student Research Assistant
Sapienza University of Rome
Jul 2017 – Oct 2018 Rome, Italy

Competition procedure: BS-J 7/2017

Extended the UCrawler framework for crawling and scraping content of research articles and citation graphs on DBLP and SemanticScholar. During this period, I also completed my master’s thesis.

Other Experience

 
 
 
 
 
Heimerer College
Assistant Professor
Heimerer College
Jul 2022 – Present Pristina, Kosovo (remote)
I’m head of the M.Sc. in Digital Healthcare. I teach Bioinformatics and AI for Health.
 
 
 
 
 
Pricewaterhouse Coopers, Digital Innovation Team
Software Engineer
Pricewaterhouse Coopers, Digital Innovation Team
Dec 2021 – Jun 2022 Rome, Italy
I focused on software prototyping and development activities. In particular, I optimized back-end services and developed highly-maintainable and efficient API services.
 
 
 
 
 
E Software Solutions
Senior Software Consultant
E Software Solutions
Sep 2020 – Mar 2021 Rome
I designed and maintained the CMS for electric vehicle leasing in the UK, Gridserve

Visiting Periods

 
 
 
 
 
Technical University of Munich
Visiting Researcher
Technical University of Munich
Jun 2023 – Sep 2023 Munich, Geramny
Visiting period at the Responsible Data Science research group with head Prof Kasneci, School of Social Sciences and Technology. Research on explainability in dynamic data landscapes and graph learning.
 
 
 
 
 
George Mason University
Visiting PhD Student
George Mason University
Apr 2021 – Jun 2021 Farifax, VA, USA
Visiting period at Prof Domeniconi’s Data Mining and Machine Learning lab. I worked on deep learning applications and anomaly detection in e-learning and learning analytics.

Honors & Awards

Sapienza University of Rome
Subject Expert
Sapienza University of Rome Feb 2024
(“Cultore della Materia” Art. 42 del R.D. 04/06/1938, n. 1269) Award for highly-skilled researcher and technician on Machine Learning course in M.Sc. Computer Science, Department of Computer Science
NeurIPS XAIA 2023 Workshop
Highlighted Reviewer
NeurIPS XAIA 2023 Workshop Dec 2023
I reviewed three workshop papers thoroughly and provided useful hints to make the submissions better. The Program Chairs emailed me that I received the Highlighted Reviewer acknowledgement. Unfortunately, I did not receive a certificate 🤨
University of L'Aquila
Subject Expert
University of L’Aquila Mar 2023
(“Cultore della Materia” Art. 42 del R.D. 04/06/1938, n. 1269) Award for highly-skilled researcher and technician on Deep Neural Networks course in M.Sc. Computer Science, Department of Information Engineering, Computer Science, and Mathematics
Sapienza University of Rome
Scholarship in AI & Computer Science
Sapienza University of Rome Nov 2021 – Sep 2022
Winner of the BS-S 6/2021 Open Competition published on September 15th 2021 (num: 1207, rep: 326, class: VII/1) on research project “000090 19 RS VELARDI - RICERCA ATENEO 2019 - MEDI PROGETTI-VELARDI (Responsabile Scientifico, VELARDI P.)”
See certificate
Sapienza University of Rome
Scholarship in AI
Sapienza University of Rome May 2020 – Apr 2021
Winner of the Avvio alla Ricerca 2020 - Tipo I, prof. Num: AR120172A8B35EEA on the research project “Personalized e-Learning Solutions to Improve the Efficacy of Learning Outcomes in Computer Science e-Courses”. I devised an autonomous model to detect students prone to drop out of university in online computer science courses, and provide them with personalized feedback and learning pathways to support their academic journey. With this money (€ 1000.00 ), I bought my PC (now pretty slow 🤣🙄) and published the paper Prenkaj et al. “Hidden Space Deep Sequential Risk Prediction on Student Trajectories”, In Future Generation Computing Systems, vol. 125, pp. 532-543, 2021.
See certificate
IRDTA
3rd International Summer School on Deep Learning
IRDTA Jul 2019 – Jul 2019
Research training event about the most recent advances in the critical and fast developing area of deep learning
LazioDiSU, Ente per il Diritto agli Studi Universitari nel Lazio
Graduation Prize (B.Sc.)
LazioDiSU, Ente per il Diritto agli Studi Universitari nel Lazio Mar 2017
Winner of the Premio di Laurea distributed from LazioDiSU to excellent bachelor degree students. Winner num: 899, grade: 110/110, Sum: € 2,559.18
See certificate
LazioDiSU, Ente per il Diritto agli Studi Universitari nel Lazio
Merit Scholarship in Computer Science
LazioDiSU, Ente per il Diritto agli Studi Universitari nel Lazio Sep 2013 – Oct 2018
Winner of the LazioDiSU Study Scholarship for B.Sc. (3 years) and M.Sc. (2 years). Yearly sum: € 5,118.36

Projects

*
All Anomaly detection Explainability Other
GRETEL
Open source framework for evaluating Graph Counterfactual Explanation (GCE) methods. Our main goal is to create a generic platform that allows the researchers to speed up the process of developing and testing new GCE methods.
PDF Code
E-Linus
Regional project - Avviso Pubblico “Emergenza Coronavirus e oltre” - Domanda prot. n. A0376-2020-070051, CUP; F84E21000000006) in support to DATAWIZARD SRL
Slides
Boosting-based Autoencoder
BAE is an unsupervised ensemble method that builds an adaptive cascade of autoencoders to achieve improved and robust results. BAE trains the autoencoder components sequentially by performing a weighted sampling of the data, aimed at reducing the amount of outliers used during training, and at injecting diversity in the ensemble.
Code
uCrawler
Universal Crawler, a concurrent Distributed and Highly Configurable and Flexible Next Generation Crawler.
Code

Invited Talks, Tutorials & Conference Presentations

Unifying Evolution, Explanation, and Discernment, A Generative Approach for Dynamic Graph Counterfactuals
I gave a talk on my paper on Dynamic Graph Counterfactuality accepted as oral presentation at KDD'24.
Oct 17, 2024 5:00 PM — 6:00 PM Online
Bardh Prenkaj
PDF Code Slides Video
Unifying Evolution, Explanation, and Discernment, A Generative Approach for Dynamic Graph Counterfactuals
Robust Stochastic Graph Generator for Counterfactual Explanations
I gave a talk on my paper on Generative Graph Counterfactuality accepted as oral presentation at AAAI'24.
Jun 15, 2024 5:00 PM — 6:00 PM Online
Bardh Prenkaj
PDF Code Slides
Robust Stochastic Graph Generator for Counterfactual Explanations
See all events

Courses

*
All Sapienza Luiss Heimerer
Machine Learning 2023/24 (24h, 6 ECTS, Laboratory classes)
PDF Code
Bioinformatics 2023/24 (30h, 6 ECTS)
PDF
Algorithms 2022/23 (26h, 8 ECTS, Laboratory classes)
PDF Code Slides Video
Machine Learning 2022/23 (24h, 6 ECTS, Laboratory classes)
PDF Code
Bioinformatics 2022/23 (30h, 6 ECTS)
PDF Slides

Recent Publications

Quickly discover relevant content by filtering publications.
Shuo Yang, Bardh Prenkaj, Gjergji Kasneci (2025). RAZOR: Sharpening Knowledge by Cutting Bias with Unsupervised Text Rewriting. In AAAI 2025.

PDF Cite Code DOI

Bardh Prenkaj, Mario Villaizan-Vallelado, Tobias Leemann, Gjergji Kasneci (2024). Unifying Evolution, Explanation, and Discernment: A Generative Approach for Dynamic Graph Counterfactuals. In KDD 2024.

PDF Cite Code Poster Slides DOI

Luca Podo, Bardh Prenkaj, Paola Velardi (2024). Agnostic Visual Recommendation Systems: Open Challenges and Future Directions. In IEEE TVCG.

Cite DOI

Mario Alfonso Prado-Romero, Bardh Prenkaj, Giovanni Stilo (2024). Robust Stochastic Graph Generator for Counterfactual Explanations. In AAAI 2024.

Cite Code Project Poster Slides DOI

Mario Alfonso Prado-Romero, Bardh Prenkaj, Giovanni Stilo, Fosca Giannotti (2023). A Survey on Graph Counterfactual Explanations: Definitions, Methods, Evaluation, and Research Challenges. In ACM CSUR.

PDF Cite Code DOI

See all publications

Contact