InterDigital and UT Austin to Showcase Over-the-Air Federated Learning at Brooklyn 6G Summit

InterDigital and UT Austin Collaborate on AI for Wireless Networks

InterDigital, Inc. (Nasdaq: IDCC), a leading research and development company in mobile, video, and AI technologies, has announced its participation in the 2024 Brooklyn 6G Summit, scheduled for October 23-25 1

1

2

2

. In collaboration with researchers from The University of Texas at Austin, InterDigital will showcase their latest advancements in AI solutions for wireless networks, particularly focusing on how AI will be utilized in 5G and 6G technologies.

Over-the-Air Federated Learning: A Novel Approach

The centerpiece of InterDigital and UT Austin's presentation will be a demonstration of Over-the-Air Federated Learning (OTA-FL), a cutting-edge technique for AI model training 1

1

. This innovative approach addresses two critical challenges in applying AI/ML to wireless networks:

1. Managing model training overhead
2. Ensuring user data privacy

Traditional federated learning requires each device to send model updates using separate dedicated radio resources to a central server. In contrast, OTA-FL aggregates model updates directly "over-the-air" 2

2

. This method offers significant advantages:

• Reduced overhead
• Improved energy efficiency, especially with an increasing number of users
• Enhanced scalability for large-scale AI model training

Live Demonstration and Real-World Applications

The joint demo will utilize software-defined radios (SDRs) to simultaneously transmit model updates, showcasing real-time OTA-FL training for image classification 1

1

. This practical demonstration highlights the potential of OTA-FL in real-world scenarios, particularly in managing the growing demands of AI in wireless networks.

Expert Insights on OTA-FL

Milind Kulkarni, VP and Head of Wireless Labs at InterDigital, emphasized the importance of new solutions for AI enablement in wireless networks, stating, "As industry prepares for the increased use of AI, new solutions are required to enable AI - especially for the training of AI models" 2

2

.

Professor Kaushik Chowdhury from UT Austin described OTA-FL as "a whole new paradigm of training where the wireless medium computes for us," highlighting its potential to handle model weight updates from numerous devices without proportionally increasing spectrum and time resources 1

1

.

Energy Efficiency in AI/ML Networks

In addition to the OTA-FL demonstration, InterDigital's Senior Principal Engineer Phil Pietraski will participate in a panel discussion on "Energy Efficiency in AI/ML Networks" 2

2

. This panel, featuring experts from NVIDIA, Fraunhofer Heinrich Hertz Institute (HHI), Rohde & Schwarz, and MediaTek, underscores the growing importance of energy-efficient AI solutions in the telecommunications industry.

Implications for 6G Development

The 2024 Brooklyn 6G Summit, jointly organized by Nokia and the NYU Wireless Research Center, brings together industry leaders to shape the future of wireless technology 1

1

. This year's theme, "6G - From vision to action," reflects the industry's focus on practical implementations of next-generation wireless technologies, with AI playing a crucial role in this evolution.