NSF Workshop held on Radio Frequency Dataset Generation, Access and Sharing. June 6- 7, 2022.
Networking researchers and professionals, across industry and academia came together for our June workshop on RFDataFactory. The workshop covered a wide range of pressing concerns regarding dataset generation, accessibility, and manageability in the field of wireless research. Key discussion topics were in the realms of:
On-going data collection drives and international priorities and rules.
Application domains for AI/ ML in wireless research and a live demonstration of a data recording API developed jointly by our team at NEU and National Instruments. Live demo was shown of the API for RF data collection on USRPs as well as of the execution of a processing pipeline for running ML based predictions on the data.
Privacy protection in RF datasets, which covered the need for clarity and easier accessibility to privacy protection rules that need to be followed with respect to RF data, and discussion in general on the concept of privacy, analyzing scenarios where RF data might inadvertently leak unintended information as well as incentives behind privacy-intrusive activities.
Dataset generation barriers covering usability and scalability of datasets, on whether we can have reproducible RF scenarios. Discussions also covered the importance of competitions in data collection and demonstrations of research progress on a focused problem using relevant datasets.
Trust and confidence building in datasets.
Data collection on PAWR platforms covering the opportunities and resources available and how more people might effectively engage and share and use RF datasets.
The workshop included a demonstration of data collection tool to collect and store data in standardized data formats as well as talk on software APIs and tools for next generation cellular networks and their potential in realizing the full potential of our networking resources.
Please find the workshop program details and video recordings of the proceedings available here.