Trustworthy AI-Enhanced Video Conferencing

Mentored by Frank Fitzek, Stefanie Speidel
at TU Dresden, ComNets Chair

In an era where digital communication predominates, traditional video conferencing systems, although effective, require the transmission of large amounts of data to maintain high-quality video (4K and beyond) and audio. This project proposes a transformative approach by minimizing the data exchanged between communication partners through advanced AI mechanisms. The core idea is to transmit only essential information such as text and emotions and to reconstruct the video and audio in real-time using sophisticated AI models, such as large language models. The project seeks to address not only technological challenges but also align with European regulatory frameworks and societal expectations.

Research Objectives:

• System Realization: Investigate architectural frameworks and methodologies to develop an AI-enhanced video conferencing system that significantly reduces data transmission while maintaining high fidelity in communication.
• Training and User Satisfaction: Determine the duration and intensity of training required for the AI models to produce outputs that satisfy real-user conditions and expectations.
• Ensuring Trustworthiness: Explore mechanisms to ensure trustworthiness in a system where the video and audio are synthetically generated. Investigate the role real video sequences can play in enhancing the authenticity and reliability of the reconstructed outputs.
• Latency Issues: Assess the latency introduced by AI processing and develop solutions to minimize its impact on real-time communication.
• Societal and Ethical Alignment: Investigate how this AI system can be developed and implemented within the existing and evolving regulatory frameworks, such as the EU’s Artificial Intelligence Act, focusing on preventing discrimination and ensuring privacy and data protection.

Work Environment

PhD candidates will be part of a dynamic research team at ComNets at TU Dresden, equipped with state-of-the-art facilities for conducting high-impact research. Candidates will have access to extensive computational resources and collaboration opportunities with leading experts in AI, machine learning, and communication technologies.

Prerequisites

• Master’s Degree (or equivalent) in Computer Science, Electrical Engineering, or a related field
• Strong programming skills (e.g., C++, Python)
• Background in one or more of the following areas:
  • AI, machine learning
  • real-time systems
  • data compression
• Ability to work effectively in an interdisciplinary team

Further details on the requirements and application process can be found in SECAI's announcement for open PhD positions in 2024.