Burkhard Stiller
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Burkhard Stiller has a total of 170 co-authors. Here are the top co-authors by number of shared publications:
- Alberto Huertas Celdrán
Communication Systems Group (CSG), the Department of Informatics (IFI), University of Zurich (UZH), Zrich, Switzerland
18 shared publications - Gérôme Bovet
Cyber-Defence Campus armasuisse Science & Technology CH-3602 Thun Switzerland
18 shared publications - Eder John Scheid
UFRGS
15 shared publications - Muriel Figueredo Franco
UFCSPA
14 shared publications - Bruno Rodrigues
University of Zurich UZH
11 shared publications - Gregorio Martínez Pérez
Department of Information and Communications Engineering University of Murcia 30100 Murcia Spain
11 shared publications
Latest Publications
- Open Access
- 7 November 2026
- Published by Association for Computing Machinery (ACM) in Distributed Ledger Technologies: Research and Practice
No Matter of Decentralization: What is Important to Users in Decentralized and Self-Sovereign Identities?
Decentralized Identity (DI) and Self-Sovereign Identity (SSI) are alternatives to traditional identity management by providing data control to users, but their adoption in the real world remains limited. Investigating what is important to users may provide meaningful insights into why adoption is slow and guide future development and deployment. Previous explorations of priorities were limited because they prioritized system qualities from the domain expert perspective, neglecting the viewpoint of end-users. This exploratory study addresses this gap by surveying potential DI/SSI system users, such as data owners ( \(N=79\) ), verifiers ( \(N=25\) ), and issuers ( \(N=35\) ), about their priorities when dealing with an identity system from a functional and problem perspective, as well as their relative priorities. The results show that data owners are more concerned about control qualities, such as consent, while verifiers are more focused on data integrity, and issuers prioritize verifiability of identity data. Importantly, when presented with a choice between competing qualities, the more reliable of the two measures used in this study, users across all roles consistently prioritized security and protection over other qualities, such as usability and decentralization. Usability and user experience (UX) were found to be of low importance to all actors when compared against other qualities, and particularly to data owners, further highlighting the need to re-evaluate trade-offs in DI/SSI systems, such as between data sovereignty and usability. Similarly, despite decentralization being the underlying feature of DI/SSI, it was found to be ranked low priority compared to other qualities for all types of users. However, a small cluster of data owners identified decentralization as a top priority, suggesting that it may be a niche quality rather than an underlying necessity. Together, the findings offer the first empirical insights into non-expert perspectives on DI/SSI technologies, mapping potential users’ priorities across actor roles to established SSI principles.- 7 November 2026
- Published by Association for Computing Machinery (ACM) in Distributed Ledger Technologies: Research and Practice
Open AccessNo Matter of Decentralization: What is Important to Users in Decentralized and Self-Sovereign Identities?
Decentralized Identity (DI) and Self-Sovereign Identity (SSI) are alternatives to traditional identity management by providing data control to users, but their adoption in the real world remains limited. Investigating what is important to users may provide meaningful insights into why adoption is slow and guide future development and deployment. Previous explorations of priorities were limited because they prioritized system qualities from the domain expert perspective, neglecting the viewpoint of end-users. This exploratory study addresses this gap by surveying potential DI/SSI system users, such as data owners ( \(N=79\) ), verifiers ( \(N=25\) ), and issuers ( \(N=35\) ), about their priorities when dealing with an identity system from a functional and problem perspective, as well as their relative priorities. The results show that data owners are more concerned about control qualities, such as consent, while verifiers are more focused on data integrity, and issuers prioritize verifiability of identity data. Importantly, when presented with a choice between competing qualities, the more reliable of the two measures used in this study, users across all roles consistently prioritized security and protection over other qualities, such as usability and decentralization. Usability and user experience (UX) were found to be of low importance to all actors when compared against other qualities, and particularly to data owners, further highlighting the need to re-evaluate trade-offs in DI/SSI systems, such as between data sovereignty and usability. Similarly, despite decentralization being the underlying feature of DI/SSI, it was found to be ranked low priority compared to other qualities for all types of users. However, a small cluster of data owners identified decentralization as a top priority, suggesting that it may be a niche quality rather than an underlying necessity. Together, the findings offer the first empirical insights into non-expert perspectives on DI/SSI technologies, mapping potential users’ priorities across actor roles to established SSI principles. - Open Access
- 3 June 2026
- Published by Elsevier in Computer Science Review
- Vol. 62
A review of internet censorship: Modern measurement and circumvention techniques
- 3 June 2026
- Published by Elsevier in Computer Science Review
- Vol. 62
Open AccessA review of internet censorship: Modern measurement and circumvention techniques
- Open Access
- 15 March 2026
- Published by Elsevier in Computer Networks
- Vol. 281
Asynchronous Cache-based Aggregation with Fairness and Filtering for Decentralized Federated Learning
- 15 March 2026
- Published by Elsevier in Computer Networks
- Vol. 281
Open AccessAsynchronous Cache-based Aggregation with Fairness and Filtering for Decentralized Federated Learning
- Top Cited PapersOpen Access
- 11 February 2026
- Published by Elsevier in Computer Networks
- Vol. 278
Design and evaluation of a robust and explainable intrusion detection framework for 5G/B5G networks
- 11 February 2026
- Published by Elsevier in Computer Networks
- Vol. 278
Top Cited PapersOpen AccessDesign and evaluation of a robust and explainable intrusion detection framework for 5G/B5G networks
- Open Access
- 31 December 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-9
FedSynthesis: A Flower-based Framework for Carbon-Reduced Federated Learning
- 31 December 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-9
Open AccessFedSynthesis: A Flower-based Framework for Carbon-Reduced Federated Learning
- Open Access
Smarter, not Bigger: Fine-Tuned RAG-Enhanced LLMs for Automotive HIL Testing
Hardware-in-the-Loop (HIL) testing is essential for automotive validation but suffers from fragmented and underutilized test artifacts. This paper presents HIL-GPT, a retrieval-augmented generation (RAG) system integrating domain-adapted large language models (LLMs) with semantic retrieval. HIL-GPT leverages embedding fine-tuning using a domain-specific dataset constructed via heuristic mining and LLM-assisted synthesis, combined with vector indexing for scalable, traceable test case and requirement retrieval. Experiments show that fine-tuned compact models, such as \texttt{bge-base-en-v1.5}, achieve a superior trade-off between accuracy, latency, and cost compared to larger models, challenging the notion that bigger is always better. An A/B user study further confirms that RAG-enhanced assistants improve perceived helpfulness, truthfulness, and satisfaction over general-purpose LLMs. These findings provide insights for deploying efficient, domain-aligned LLM-based assistants in industrial HIL environments.
Open AccessSmarter, not Bigger: Fine-Tuned RAG-Enhanced LLMs for Automotive HIL Testing
Hardware-in-the-Loop (HIL) testing is essential for automotive validation but suffers from fragmented and underutilized test artifacts. This paper presents HIL-GPT, a retrieval-augmented generation (RAG) system integrating domain-adapted large language models (LLMs) with semantic retrieval. HIL-GPT leverages embedding fine-tuning using a domain-specific dataset constructed via heuristic mining and LLM-assisted synthesis, combined with vector indexing for scalable, traceable test case and requirement retrieval. Experiments show that fine-tuned compact models, such as \texttt{bge-base-en-v1.5}, achieve a superior trade-off between accuracy, latency, and cost compared to larger models, challenging the notion that bigger is always better. An A/B user study further confirms that RAG-enhanced assistants improve perceived helpfulness, truthfulness, and satisfaction over general-purpose LLMs. These findings provide insights for deploying efficient, domain-aligned LLM-based assistants in industrial HIL environments.
- Open Access
- 18 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 26-33
Bridging Technical Capability and User Accessibility: Off-grid Civilian Emergency Communication
- 18 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 26-33
Open AccessBridging Technical Capability and User Accessibility: Off-grid Civilian Emergency Communication
- Open Access
- 18 October 2025
- Published by Elsevier in Information and Software Technology
- Vol. 190
GreenDFL: A framework for assessing the sustainability of Decentralized Federated Learning systems
- 18 October 2025
- Published by Elsevier in Information and Software Technology
- Vol. 190
Open AccessGreenDFL: A framework for assessing the sustainability of Decentralized Federated Learning systems
- Open Access
- 24 August 2025
- Published by Association for Computing Machinery (ACM)
- p. 682-695
QUIC-Exfil: Exploiting QUIC's Server Preferred Address Feature to Perform Data Exfiltration Attacks
- 24 August 2025
- Published by Association for Computing Machinery (ACM)
- p. 682-695
Open AccessQUIC-Exfil: Exploiting QUIC's Server Preferred Address Feature to Perform Data Exfiltration Attacks
- Open Access
FEST: A Unified Framework for Evaluating Synthetic Tabular Data
Synthetic data generation, leveraging generative machine learning techniques, offers a promising approach to mitigating privacy concerns associated with real-world data usage. Synthetic data closely resembles real-world data while maintaining strong privacy guarantees. However, a comprehensive assessment framework is still missing in the evaluation of synthetic data generation, especially when considering the balance between privacy preservation and data utility in synthetic data. This research bridges this gap by proposing FEST, a systematic framework for evaluating synthetic tabular data. FEST integrates diverse privacy metrics (attack-based and distance-based), along with similarity and machine learning utility metrics, to provide a holistic assessment. We develop FEST as an open-source Python-based library and validate it on multiple datasets, demonstrating its effectiveness in analyzing the privacy-utility trade-off of different synthetic data generation models. The source code of FEST is available on Github.
Open AccessFEST: A Unified Framework for Evaluating Synthetic Tabular Data
Synthetic data generation, leveraging generative machine learning techniques, offers a promising approach to mitigating privacy concerns associated with real-world data usage. Synthetic data closely resembles real-world data while maintaining strong privacy guarantees. However, a comprehensive assessment framework is still missing in the evaluation of synthetic data generation, especially when considering the balance between privacy preservation and data utility in synthetic data. This research bridges this gap by proposing FEST, a systematic framework for evaluating synthetic tabular data. FEST integrates diverse privacy metrics (attack-based and distance-based), along with similarity and machine learning utility metrics, to provide a holistic assessment. We develop FEST as an open-source Python-based library and validate it on multiple datasets, demonstrating its effectiveness in analyzing the privacy-utility trade-off of different synthetic data generation models. The source code of FEST is available on Github.