2nd RESTART Tech Camp on 5G and Open RAN | September 9-12, Milan (Italy)

This is the repository for the O-RAN assignment/hands-on related to the tech camp.

You will learn to work with cellular datasets and train AI solutions to classify traffic profiles.

• The script Train_Predict_Light.py offers a baseline training pipeline that shows you how to open the dataset, extract/plot data, and provide a skeleton to train a base AI algorithm.

• The script Test_Performance_Fake.py offers a testing algorithm for your trained algorithm to verify that your solution processes data correctly and can be evaluated.

• The script Train_Predict_Full.ipynb offers a jupyter notebook containing a baseline training pipeline that shows you how to open the dataset, extract/plot data, and provide you an implementation of a traditional training & prediction AI pipeline for two classifiers of example.

The datasets you need to download are located at:

• Fake testing dataset available at this link. This has the same format as the testing dataset but has not the same data that will be used to evaluate your algorithms :) We will make this dataset available once you are all done with the activities.
• Training dataset available at this link. This is the training dataset you should use. We suggest you split this into 3 subsets for training, validation and testing. Note that the testing dataset you will produce here is different from the one above.

Important:

• We will ask you to submit a trained model we can test using the ./Test_Performance_Fake.py script. If you cannot test your model using that script, it means that you need to change input/output format to match the requirement.
• Groups that are willing to present their approach, are free to prepare 2-3 slides to present their work and results to the audience. Please let us know in advance if you are willing to present.

Slides from the talk:

• ML Primer: this link

• Detaset description slides: this link

• Original dataset available at this link

• Paper describing the dataset and AI-based xApps is M. Polese, L. Bonati, S. D'Oro, S. Basagni, T. Melodia, "ColO-RAN: Developing Machine Learning-based xApps for Open RAN Closed-loop Control on Programmable Experimental Platforms," IEEE Transactions on Mobile Computing, pp. 1-14, July 2022. PDF

  **Email contacts

  • Andrea Pimpinella: andrea.pimpinella@unibg.it
  • Salvatore D'Oro: s.doro@northeastern.edu