Guest lecture Thursday March 7th, 2024: DPhil student Changgang Zheng, Engineering Science at the University of Oxford

The lecture will be held in person at Lille Aud (B203) in the Science building (Realfagsbygget).

Professional (academic) contact person is Post doctor Masoud Hemmatpour.

Target group: PhD students and employees at CS department.

Welcome!

----------------

Date: Thursday March 7th, 2024, at 14:15
Place: Lille Aud (B203) in the Science building (Realfagsbygget)
Title: " Toward Distributed In-Network Machine Learning "
Lecturer: DPhil student Changgang Zheng, Engineering Science at the University of Oxford

Abstract
Deploying in-network machine learning applications in programmable network devices faces challenges in scalability due to resource constraints. This presentation introduces DINC, a distributed deployment framework used to scale in-network computing models. The framework generates deployment strategies, overcomes resource constraints and provides functionality guarantees across a network. DINC uses multi-objective optimization to provide a deployment strategy, slicing P4 programs accordingly. DINC was evaluated using seven different workloads on both data center and wide-area network topologies, demonstrating feasibility and scalability, providing efficient distribution plans within seconds. DINC is open-source and publicly available via GitHub.

Short bio
I am Changgang Zheng, a final-year DPhil student in Engineering Science at the University of Oxford. My research interests include networking, in-network computing, and machine learning. A specific interest of mine is applying in-network computing and machine learning techniques to address networking challenges.

Kortnytt fra Institutt for informatikk

Fuglesteg, Jan jan.fuglesteg@uit.no Studierådgiver