(Untitled)
agenda-interim-2025-nmrg-01-nmrg-01-02

NMRG 80th meeting
Online

Tuesday 2025-01-21 11:00-13:30 CET

Contacts:
- RG Chairs
- Laurent Ciavaglia <laurent.ciavaglia@nokia.com>
- Jérôme François <jerome.francois@uni.lu>
- RG Secretaries
- Jéferson Campos Nobre <jcnobre@inf.ufrgs.br>
- Pedro Martinez-Julia <pedromj@gmail.com>

Useful links:
- Materials: https://datatracker.ietf.org/doc/agenda-interim-2025-nmrg-01-nmrg-01/
- Meetecho: https://meetings.conf.meetecho.com/interim/?group=3905ec74-118f-46ef-acb7-d2ca8202bcec
- Notes: https://notes.ietf.org/notes-ietf-interim-2025-nmrg-01-nmrg
- Video recording: https://www.youtube.com/user/ietf/playlists (available post-meeting)

** Agenda **

1. Introduction, RG Chairs, 5 min

2. Data descriptors and topologies in decentralised learning, Arashmid Akhavain, 20 min including Q&A

The model-follow-data paradigm employed by decentralised learning methods trains AI models by moving them to data nodes. It is imperative to be able to identify nodes with suitable data for each model in these knowledge sharing networks. This presentation explores the necessity and potential requirements of data descriptors and topologies as a vehicle for achieving the aforementioned objective.

3. Characterizing NIDS Datasets: From Bad Design Smells Towards Actionable Metrics, Kokouvi Benoit Nougnanke, 25 min including Q&A

4. Classification with Synthetic Radio Data for Real-life Environment Sensing, Marie Line, 25 min including Q&A

Largescale qualitative data is required for achieving high accuracy learning. However, their training phase leads to prohibitive cost and heavy constraints on data collection and storage that are not desirable for network. To overcome this problem, we propose to use synthetic data instead of real data for training machine learning models to avoid high cost data sharing/storage. In this paper, we are interested in real-life Environment Sensing Network in a context of limited data amount sharing. We focus on unsupervised machine learning classification models for Context Detection. A comparative study of four well-known generative models, that are able to generate synthetic 3GPP radio data with similar distribution than the source data. We investigate the quality of these synthetic generated radio data according to three dimensions: distribution similarity, data variability and detection capability. The classification models trained with synthetic generated data are tested in real-life context to infer whether a user connected to the network is inside or outside a building.

5. Chair's wrap-up on NMRG activity regarding the "data" topic, 10 min

6. Open discussion, 1h