RFDataFactory

• 9.10am - 9.15am: Goal setting, timeline, workshop overview

Kaushik Chowdhury, WIOT/Northeastern University

• 9.15am - 10.00am: Keynote - I : Nada Golmie*, NIST Fellow and Chief, Wireless Networks Division at NIST

Title: RF datasets: Seeking wireless gold

Abstract: While applying ML and AI to solving wireless communications problems have become standard practice in the telecommunication industry due to the abundance of communication data, curated RF datasets remain scarce for R&D. In this talk, we reflect on the state of the art and provide examples and lessons learned from experience in disseminating high-quality measurement datasets.

• 10.30am -11.15am: Informational Session I International priorities and rules for RF data collection

1. Joerg Widmer*, IMDEA Networks, Spain

2. Falko Dressler*, Technische Universität Berlin, Germany

3. Berk Canberk*, Istanbul Technical University, Turkey

Question 1. Large-scale efforts in RF dataset-related activities external to the US

Question 2. Identifying international partners for collaboration, data access rules

Question 3. Interception laws, privacy enforcing methods

• 1.30pm-2.15pm: Discussion Panel I. Technical architectures for privacy protection in RF datasets

1. Anant Sahai, UC Berkeley (moderator)

2. Stratis Ioannidis, Northeastern University

3. Dola Saha, SUNY Albany

4. Charles Schroeder, National Instruments

Question 1. Defining "privacy", given ubiquitious RF signals, legal implications

Question 2. Cost/benefits of privacy protection at the source versus dissemination point

Question 3. Methods of privacy protection

• 4.00pm-4.45pm: Discussion Panel II. Dataset generation barriers: software and human factors

1. Monisha Ghosh, University of Notre Dame (moderator)

2. Fatemeh Afgah, Clemson University

3. Geoffrey Mainland, Drexel University

4. Vijay Mishra, United States Army Research Laboratory

Question 1. State of the art and gaps in software tools and APIs for data collection

Question 2. Can we standardize metadata representation for diverse RF use-cases?

Question 3. Ways to encourage community engagement

============

• 8.15am-9.00am Workshop Breakfast

• 9.00am-9.45am: Keynote - II: Lisa Guess, FCC, Technological Advisory Committee

Title: The Importance of Artificial Intelligence and Data for the Telecommunications Industry

Abstract: While the term “Artificial Intelligence” is identified with many different methods, techniques, and specialized application areas and evokes the not un-common perceptions of computers being able to exhibit human traits (referred to as General AI), it is important to our industry to take a focused view of AI specific to telecommunications and how AI can help in very tactical, near-term ways. AI methods create value by performing tasks that are hard for human beings to do, including the automation of routine networking processes, and the ability to handle large numbers of simple transactions and data sets at scale. This talk will focus on a few specific aspects of AI, including the data governance and curation of RF data sets, the sources of data, and the needs of the overall Network ecosystem.

• 10.30am -11.15am: Informational Session II. At-scale data collection through PAWR platforms

1. Ashu Sabharwal, Rice University (moderator)

2. Magreth Mushi, NCSU

3. Sneha Kasera, University of Utah

4. Yong Guan, Iowa State University

Question 1. Existing opportunities for dataset generation, successful examples

Question 2. Lowering the barrier for new and first-time users

Question 3. Storing, curating, disseminating datasets; datasets as a service

• 1.10pm-2.00pm: Discussion Panel III. Building public trust and confidence

1. Francesco Restuccia, WIOT/Northeastern University (moderator)

2. Shiwen Mao, Auburn University

3. Mariya Zheleva, SUNY Albany

4. Nicholas Laneman, University of Notre Dame

Question 1. Allowing third party verification, novel "certifiation authorities"

Question 2. Metrics for endorsing high-value, re-usable and well designed datasets

Question 3. Topics for open-to-public competitions using high-value datasets

• 2.00pm-3.00pm: Small group breakouts III

Group 1 Zoom link /Room 136 [discussion co-lead: Kai Li]

Group 2 Zoom link /Room 138 [discussion co-lead: Brian Kim]

============

• 3.00pm-3.30pm: --------Networking break--------

============

• 3.30pm-3.45pm: Report on breakout III

• 3.45pm-4.00pm: Summary of key takeaways and action items

============

• 4.00pm: Concluding statements by NSF [Workshop adjourned]