Autonomous or self-driving networks are expected to provide a solution to the myriad of extremely demanding new applications in the Future Internet. The key to handle complexity is to perform tasks like network optimization and failure recovery with minimal human supervision. For this purpose, the community relies on the development of new Machine Learning (ML) models and techniques. However, ML can only be as good as the data it is fitted with. Datasets provided to the community as benchmarks for research purposes, which have a relevant impact in research findings and directions, are assumed to be of good quality by default, but often they are not.

In the scope of autonomous networks, assuring the quality of the data gathered by the network telemetry framework (RFC 9232) is a major need. Autonomous networks require a suitable data generation mechanism, and this should ensure that high quality data is produced. As a result, data quality assessment has to be embedded in the automatic generation of data, which calls for new approaches to assess data quality and to produce data that maximizes quality (data-quality-by-design). As for today, the process of data collection usually does emphasize quality metrics and assessment, which may lead to issues during the use. To give some simplistic examples, the data to estimate a traffic matrix to, e.g., optimize routing can be gathered with different technologies, but not all of them have the same tradeoff between accuracy and resource consumption, and the choice can affect the result of the optimization. On the other hand, a dataset for anomaly detection and diagnosis/troubleshooting often needs to be properly labelled, and inaccuracies in the labelling process can dramatically affect the performance of AI for anomaly detection.

The workshop calls for papers fitting into the following requirements:

* The paper needs to explain the data, including measurement means and the application for which the data can be useful. These datasets are expected to become benchmarks in the context of a specific problem, so this point is important.

* The data must be made available for the community.

* The data should have gone through a process of anonymization following the state-of-the-art measures, which must be explained in the paper.

* The paper is required to include a part in which the quality of the data is assessed. This will push the development of new techniques for data quality assessment. Data quality is a very transversal topic, and depending on the application the data is used for, the assessment may be done in a different way.

* It is highly recommended to provide a detailed description of the architecture employed to capture the data and the corresponding measurement process utilized. The details must allow for the replication of the environment to produce future datasets.

* Replicability papers based on already available public data sets are also in the scope of the workshop. These submissions must include the prefix 'Replication:' in their title.


Important Dates:

Paper Submission Deadline: August 25, 2025
Acceptance Notification: September 15, 2025
Camera Ready: September 29, 2025

Paper Submission:
Papers should be prepared using the IEEE 2-column conference style and are limited to 6 pages including references (full papers) or 4 pages including references (short papers). Papers must be submitted electronically in PDF format through the EDAS system.

Submission link: https://edas.info/N34165