Title | AI/ML Models for NFV Usecases (Research and Develop) |
Status | |
Difficulty |
|
Description
This project aims to develop AI/ML models for NFV-usecases. Any two of the following three problems can be considered.
VNF/CNF resource/performance/failure prediction
Problem | Model | Link | Comments |
|---|---|---|---|
Prediction of VNF Resource Demands | RNN, LSTM | https://ieeexplore.ieee.org/document/8806620 | |
NFV log analysis with NLP | BERT | ||
Synthetic monitoring and logging data generation using GANs | GAN,CycleGAN,SeqGAN |
LFN Thoth: https://wiki.anuket.io/display/HOME/Thoth
LFN Acumos: https://www.acumos.org/
TensorFlow Time Series : https://www.tensorflow.org/tutorials/structured_data/time_series
Collectd : https://collectd.org/
Repo: https://github.com/opnfv/thoth
ML Techniques: Deep_learning.
ML model development
AI/ML for Telco Usecases.
Develop AI/ML model for NFV/Telco Usecases.
Deploy model with Acumos/Kubeflow framework.
Run Acumos/Kubeflow in Anuket Testbeds.
Comprehensive report on applications of AI/ML in Networking(Comparative analysis).
Anuket Thoth
At least undergraduate
Knowledge of ML and ML-Tools - Tensorflow.
This work can be enhanced to more useful and complex AI for NFV usecases in future.
Part-Time
Please read all instructions before applying. Include Resume, proof of school enrollment, and participation permission from school/employer
Lei Huang <huangleiyjy@chinamobile.com> (Lei Huang)
Sridhar Rao <srao@linuxfoundation.org> (Sridhar Rao)