Deep learning and Artificial intelligence for High-Performance Networks


This project is supported by DOE ASCR Early Career Award.

With advances in computing, data is being produced at exponential rates requiring highly flexible mobility across HPC computing and distributed facilities. Networks are the essential bloodline to science collaborations across the globe such as in high-energy physics, earth sciences and genomics. However, upgrading network hardware, with high-end routers and optic fibers, to cope with this data revolution can cost millions of dollars. In this project, we are exploring artificial intelligence to design and efficiently manage distributed network architectures to improve data transfers, guarantee high-throughput and improve traffic engineering.

In this team, we are doing research in machine (and deep) learning, control engineering, unsupervised machine learning, parallel computing and network research. Our issues are guided by problems arising in wide area networks (WAN), sensor or IoT networks, distributed computing and quantum networking challenges.

Main challenges being addressed:

  1. Network Automation: Service-aware network orchestration and deployment.
  2. Traffic Engineering: Improving data transfer and network utilization via traffic bursts prediction and intelligent network controllers.
  3. Investigating deep reinforcement learning research for networking challenges.


For more information, contact Mariam Kiran <mkiran@es.net>

Deep Time-Series Prediction:

Techniques in Deep Learning, Fourier Transforms, etc, for Network Traffic prediction

Classification in Fast-High- Speed Telemetry Network Big Data:

Identifying packet loss in TCP flows for improving Network Performance

Deep Reinforcement Learning for Routing:

Developing algorithms such as DQN, DDPG and more for optimal control on networks

Intelligent Distributed Data Movement across Multi-Domains:

Optimize end-to-end network performance for distributed science workflows

Intent-based Networking

High-level language for automation of network configurations

IoT Sensors and Networks

Improving and reliable data movement from sensors

Distributed and Edge AI:

Deploying Distributed AI solutions for Facilities

Quantum Networking Research

Exploring distributed quantum systems