26617 research outputs found
Sort by
Towards an adaptive, trustworthy and privacy preserving federated learning system
No full textTOWARDS AN ADAPTIVE, TRUSTWORTHY AND PRIVACY PRESERVING FEDERATED LEARNING SYSTEM
Towards an adaptive, trustworthy and privacy preserving federated learning system (1)
Dedication (ii)
Acknowledgments (iii)
List of Figures (vii)
List of Tables (viii)
List of Abbreviations (ix)
1 Introduction (2)
2 Dynamic Behavior Assessment Protocol for Secure Decentralized Federated Learning (4)
2.1 Introduction (4)
2.2 Related Work (5)
2.3 Behavior-Based Attack Vectors in DFL Systems (9)
2.4 Proposed DFL Architecture (10)
2.4.1 Decentralized Federated Learning (11)
2.4.2 Threat Model (12)
2.4.3 Behavior Evaluation (12)
2.4.4 Dynamic Reputation Assessment (13)
2.4.5 Privacy Preserving Model (15)
2.5 Experiments and Results (15)
2.5.1 Experimental Setup (15)
2.5.2 Results and Discussion (16)
2.6 Limitations (19)
2.7 Conclusion (21)
3 OpenFL: A Scalable and Secure Decentralized Federated Learning System on the Ethereum Blockchain (22)
3.1 Introduction (22)
3.2 Background Work (24)
3.3 Related Work (24)
3.4 System Overview (26)
3.4.1 Process Description (27)
3.4.1.1 Deployment (27)
3.4.1.2 Registration (28)
3.4.1.3 Contribution and Exit (28)
3.4.2 Federated Learning Rounds (29)
3.4.2.1 Local Training (29)
3.4.2.2 Fingerprinting Weights (30)
3.4.2.3 Evaluation and Feedback (30)
3.4.2.4 Round Settlement (32)
3.4.2.5 Merging (32)
3.5 Identified Attack Vectors (34)
3.5.1 Reluctant Contribution (34)
3.5.2 False Reward Claiming (35)
3.5.3 Detrimental Contribution (35)
3.5.4 Weight or Feedback Withholding (36)
3.5.5 Contract Governance Take-Over (36)
3.5.6 Weight Front-Running (37)
3.6 Experiments (38)
3.7 Limitations (41)
3.8 Conclusion (43)
4 Split Averaging in Federated Learning: Bridging the Heterogeneity Gap in Clients Data Distributions (44)
4.1 Introduction (44)
4.2 Related Work (46)
4.3 Background and Problem Statement (48)
4.3.1 Federated Learning (48)
4.3.2 Heterogeneity in FL (49)
4.3.2.1 LABEL DISTRIBUTION SKEW (49)
4.3.2.2 FEATURE DISTRIBUTION SKEW (49)
4.3.3 Cross Industrial Data Collaboration (50)
4.3.4 The Problem Statement (51)
4.4 System Model (51)
4.5 Results (53)
4.5.0.1 EXPERIMENTAL SETUP (53)
4.5.0.2 PERFORMANCE EVALUATION (53)
4.6 Discussion and Limitations (56)
4.7 Conclusion (57)
5 Adaptive Privacy-Preserving Federated Learning for Robust IoT Systems: A Defense Against Data Poisoning Attacks (58)
5.1 Introduction (58)
5.2 Related Work (59)
5.3 Core Technologies (61)
5.3.1 Differential Privacy (61)
5.3.2 Federated Learning (62)
5.3.3 Motivational Scenarios: (63)
5.3.3.1 Cross Industry Data Collaboration (63)
5.3.3.2 Model Exchange Between Renewable Energy Plants. (63)
5.4 Adaptive and Privacy-Preserving FL Method (64)
5.4.1 APPFL Method (64)
5.4.2 Determining Optimal Clients Weights (65)
5.5 Experimental Setup (65)
5.6 Results (67)
5.7 Discussion and Limitations (71)
5.8 Conclusion (73)
6 Future Research Directions (74)
6.1 Federated Unlearning (74)
6.2 Explainable Federated Learning (74)
6.3 Zero or Few Shot Federated Learning (74)
6.4 Federated Transfer Learning (74)
6.5 Privacy Risks Due to Generative AI (75)
7 Conclusions (76)
References (76
Leveraging GenAI for innovation-driven advantage
No full textLEVERAGING GENAI FOR INNOVATION-DRIVEN ADVANTAGE
Leveraging GenAI for innovation-driven advantage (1)
List of Figures (viii)
List of Tables (viii)
Abbreviation List (ix)
1. Introduction (1)
1.1. Background (1)
1.2. Problem Statement and Research Questions (2)
1.2.1. Research Questions (4)
1.2.2. Research Sub-Questions (4)
1.3. Relevance of the Thesis (5)
1.3.1. Theoretical Relevance (5)
1.3.2. Practical Relevance (5)
1.3.3. Target Audience (5)
1.4. Objective and Structure of the Thesis (6)
2. Literature Review (7)
2.1. Key Concepts (7)
2.1.1. Artificial Intelligence and Generative AI (7)
2.1.2. Innovation (9)
2.1.3. Sustainable Competitive Advantage (11)
2.2. Theoretical Framework (13)
2.2.1. Resource-Based View (13)
2.2.2. Dynamic Capabilities (14)
2.3. Overview of Relevant Literature (15)
2.3.1. From Generative AI to Sustainable Competitive Advantage (15)
2.3.2. The Role of Generative AI in Enhancing Innovation Processes (18)
2.3.3. Innovation as a Driver of Sustainable Competitive Advantage (20)
2.3.4. Broader Considerations Shaping GenAI and Sustainable Competitive Advantage (22)
2.4. Research Gap (24)
3. Methodology (26)
3.1. Research Design and Justification (26)
3.2. Interview Development and Guideline (27)
3.3. Data Collection Procedure (28)
3.3.1. Sampling and Participant Selection (28)
3.3.2. Participants (29)
3.3.3. Interview Procedure (31)
3.4. Data Analysis Procedure (32)
4. Results (33)
4.1. Presentation of Findings (33)
4.1.1. Overview of Interviewed Firms and Use Cases (33)
4.1.2. Types of GenAI Tools Used Internally (35)
4.1.3. Objectives Behind GenAI Adoption (37)
4.1.4. GenAI Integration into Innovation and Operations (38)
4.1.5. Enabling Factors of GenAI Use (39)
4.1.6. Challenges Encountered in GenAI Implementation (40)
4.1.7. Perceived Outcomes and View on Sustainable Competitive Advantage (42)
4.2. Data Analysis and Interpretation (44)
4.2.1. Patterns in Adoption Motivations and Strategic Intent (44)
4.2.2. Role of Organisational Context in Shaping GenAI Use (45)
4.2.3. Human-AI Collaboration and Workforce Impact (45)
4.3. Additional Insights (46)
4.3.1. Forward-Looking Views (46)
4.3.2. Practical Recommendations (48)
4.3.3. Broader Considerations for Responsible Use (49)
5. Discussion (50)
5.1. Interpretation of Results (50)
5.2. Theoretical and Practical Implications (54)
5.2.1. Theoretical Implications (54)
5.2.2. Practical Implications (54)
5.3. Limitations (55)
5.4. Future Research Directions (57)
6. Conclusion (58)
7. References (60)
8. Appendices (71)
8.1. Interview Guide (71)
8.2. Interview Transcripts (74)
8.2.1. Interview I1: 23 May 2025, 12:00 via Microsoft Teams (74)
8.2.2. Interview I2: 28 May 2025, 10:00 via Microsoft Teams (82)
8.2.3. Interview I3: 2 June 2025, 10:30 via Microsoft Teams (90)
8.2.4. Interview I4: 5 June 2025, 11:00 via Microsoft Teams (102)
8.2.5. Interview I5: 6 June 2025, 10:30 via Microsoft Teams (111)
8.2.6. Interview I6: 6 June 2025, 14:30 via Microsoft Teams (122)
8.2.7. Interview I7: 13 June 2025, 09:00 via Microsoft Teams (131)
8.3. Coded Findings (142)
8.3.1. Interview I1 – Interview Code (142)
8.3.2. Interview I2 – Interview Code (147)
8.3.3. Interview I3 – Interview Code (152)
8.3.4. Interview I4 – Interview Code (159)
8.3.5. Interview I5 – Interview Code (164)
8.3.6. Interview I6 – Interview Code (170)
8.3.7. Interview I7 – Interview Code (176
Getränkeverpackungen im Wandel?
No full textGETRÄNKEVERPACKUNGEN IM WANDEL?
Getränkeverpackungen im Wandel? (1
Creating public value through AI use in high-risk services
No full textCREATING PUBLIC VALUE THROUGH AI USE IN HIGH-RISK SERVICES
Creating public value through AI use in high-risk services (1)
Abstract (4)
1 Introduction (9)
2 Theoretical Foundations (12)
2.1 Understanding Public Value in Public Services (12)
2.1.1 The Foundations of Public Value (12)
2.1.2 The Strategic Triangle (13)
2.1.2.1 Legitimacy and Support (14)
2.1.2.2 Public Value Creation (15)
2.1.2.3 Operational Capacity (16)
2.1.3 Risk Aversion as a Barrier to Innovation (16)
2.2 AI in High-Risk Public Services (17)
2.2.1 Task Automation in the Hands of AI (18)
2.2.2 The Collaborative Paradigm (19)
2.3 Challenges of AI Applications in Public Services (20)
2.3.1 Trust and Public Value (20)
2.3.2 Algorithmic Reliance (21)
2.3.3 Real-World Failures of AI in Public Services (23)
2.3.4 Operational Limitations (24)
3 Methodology (26)
3.1 Desk Research (26)
3.2 Refined Search Terms (26)
3.3 Research Design (28)
3.4 Survey Design and Protocol (28)
3.4.1 Sample (31)
3.5 Interview Design and Protocol (32)
3.5.1 Structure and Format (33)
3.6 Data Analysis (34)
3.6.1 Quantitative Analysis (34)
3.6.2 Qualitative Analysis (35)
4 Analysis of Quantitative Results (37)
4.1 Public Value Perceptions (37)
4.2 Operational Capacity Perceptions (37)
4.3 Legitimacy & Support Perceptions (38)
4.4 Mediating Influences on AI Perceptions (41)
4.4.1 Trust in Government (41)
4.4.2 Frequency of AI Usage (41)
4.4.3 Gender Differences (42)
4.4.4 Other Factors: Age and Education (43)
5 Analysis of Qualitative Results (44)
5.1 Open-ended survey questions (44)
5.1.1 Privacy & Surveillance Concerns (44)
5.1.2 Bias and Discrimination (44)
5.1.3 Transparency and Accountability (45)
5.1.4 Human Involvement and Empathy (46)
Across all scenarios, participants put a strong emphasis on the importance of human judgement as the final step in AI-driven decision making. Especially in sensitive issues which impact the individual directly such as triage, law enforcement and immig... (46)
5.1.5 Operational Challenges (46)
5.2 Interviews (47)
5.2.1 Sensitivity of the Decision Context as a Key Driver of Trust (47)
5.2.2 Transparency and Accountability as Essential Prerequisites (48)
5.2.3 Administrative and Internal Challenges (49)
6 Discussion (50)
6.1 Empirical Reflections on Moore’s Triangle (50)
6.1.1 Public Value Creation (50)
6.1.2 Legitimacy and Support (52)
6.1.3 Operational Capacity (54)
6.1.4 Moore’s Strategic Triangle Reframed for AI-Driven Public Services (55)
6.1.4.1 Public Value: Citizen Perceptions of Value, Fairness, Equity (56)
6.1.4.2 Operational Capacity: Technical Readiness, AI Literacy, Human Oversight (56)
6.1.4.3 Legitimacy & Support: Institutional Trust, Transparency, Accountability (57)
6.2 Factors Influencing Perceptions of Public Value (57)
6.2.1 Fairness and Equity (58)
6.2.2 Transparency and Explainability (58)
6.2.3 Human Oversight and Accountability (59)
6.2.4 Trust in Government (59)
6.2.5 Frequency of AI Use (60)
7 Conclusion (61)
7.1 Contributions to Theory (61)
7.2 Contributions to Practice (61)
7.3 Limitations (62)
8 References (63)
9 Appendix (70)
9.1 Appendix A: Survey Questionnaire (70)
9.2 Appendix B: Interview Guideline (74)
9.3 Appendix C: Coding Scheme for Open-ended Survey Questions (75)
9.4 Appendix D: Coding Scheme for the Interviews (76
What role does user trust play in resolving the privacy-personalization-performance paradox in technologies?
No full textWHAT ROLE DOES USER TRUST PLAY IN RESOLVING THE PRIVACY-PERSONALIZATION-PERFORMANCE PARADOX IN TECHNOLOGIES?
What role does user trust play in resolving the privacy-personalization-performance paradox in technologies? (1
Berücksichtigung und Quantifizierung des Insolvenzrisikos bei der Bewertung von Startups
No full textBERÜCKSICHTIGUNG UND QUANTIFIZIERUNG DES INSOLVENZRISIKOS BEI DER BEWERTUNG VON STARTUPS
Berücksichtigung und Quantifizierung des Insolvenzrisikos bei der Bewertung von Startups (1
Strategic and intuitive drivers of role differentiation in early-stage tech startups
No full textSTRATEGIC AND INTUITIVE DRIVERS OF ROLE DIFFERENTIATION IN EARLY-STAGE TECH STARTUPS
Strategic and intuitive drivers of role differentiation in early-stage tech startups (1
Exploring the dynamics of buyer-supplier relationship termination in global value chains
No full textEXPLORING THE DYNAMICS OF BUYER-SUPPLIER RELATIONSHIP TERMINATION IN GLOBAL VALUE CHAINS
Exploring the dynamics of buyer-supplier relationship termination in global value chains (1)
I. Abstract (2)
II. List of Abbreviations (4)
III. List of Figures (5)
IV. List of Tables (5)
1 Introduction (6)
2 Theoretical Background (9)
2.1 Global Value Chains (GVCs) (9)
2.1.1 Defining Global Value Chains (9)
2.1.2 Evolution and Expansion of GVCs (GVCs structure before) (10)
2.1.3 Characteristics and Structure of GVCs today (14)
2.2 Buyer-Supplier Relationship (16)
2.2.1 Definition and Importance of Buyer-Supplier Relationship (16)
2.2.2 Buyer-Supplier Relationships typologies (17)
2.2.3 Trust (17)
2.2.4 Power-dependance and Ressource Dependance Theory (18)
2.3 Corporate Social Responsibility and Irresponsibility (23)
2.3.1 Corporate Social Responsibility (CSR) (23)
2.3.2 Corporate Social Irresponsibility (CSI) (26)
2.3.3 Causes of Corporate Social Irresponsibility (CSI) (28)
3 Hypothesis Development (30)
3.1 Justification of Hypothesis H1 (31)
3.2 Justification of Hypothesis H2 (37)
4 Methodology (42)
4.1 Sample (43)
4.2 Dependent variable (44)
4.3 Independent variable (45)
4.4 Moderator (46)
4.5 Control Variables (46)
4.6 Econometric model (47)
5 Empirical Results (47)
6 General Discussion and Implications (51)
6.1 Theoretical Implications (51)
6.2 Practical Implications (53)
6.3 Limitations (55)
7 References (57
Die richterliche Unparteilichkeit in der öffentlich-rechtlichen Gerichtsbarkeit
No full textDIE RICHTERLICHE UNPARTEILICHKEIT IN DER ÖFFENTLICH-RECHTLICHEN GERICHTSBARKEIT
Die richterliche Unparteilichkeit in der öffentlich-rechtlichen Gerichtsbarkeit (1