Lost in Transmission: On the Impact of Networking Corruptions on Video Machine Learning Models

06/10/2022
by   Trenton Chang, et al.
0

We study how networking corruptions–data corruptions caused by networking errors–affect video machine learning (ML) models. We discover apparent networking corruptions in Kinetics-400, a benchmark video ML dataset. In a simulation study, we investigate (1) what artifacts networking corruptions cause, (2) how such artifacts affect ML models, and (3) whether standard robustness methods can mitigate their negative effects. We find that networking corruptions cause visual and temporal artifacts (i.e., smeared colors or frame drops). These networking corruptions degrade performance on a variety of video ML tasks, but effects vary by task and dataset, depending on how much temporal context the tasks require. Lastly, we evaluate data augmentation–a standard defense for data corruptions–but find that it does not recover performance.

READ FULL TEXT

page 2

page 5

page 6

page 7

page 8

page 9

page 11

page 29

research
06/03/2019

The Adversarial Machine Learning Conundrum: Can The Insecurity of ML Become The Achilles' Heel of Cognitive Networks?

The holy grail of networking is to create cognitive networks that organi...
research
10/21/2022

Management of Machine Learning Lifecycle Artifacts: A Survey

The explorative and iterative nature of developing and operating machine...
research
04/29/2019

Challenges and Pitfalls of Reproducing Machine Learning Artifacts

An increasingly complex and diverse collection of Machine Learning(ML) m...
research
10/27/2021

Towards Intelligent Load Balancing in Data Centers

Network load balancers are important components in data centers to provi...
research
06/19/2023

Modular Simulation Environment Towards OTN AI-based Solutions

The current trend for highly dynamic and virtualized networking infrastr...
research
09/18/2023

Contrastive Learning and Data Augmentation in Traffic Classification Using a Flowpic Input Representation

Over the last years we witnessed a renewed interest towards Traffic Clas...
research
10/17/2022

Confound-leakage: Confound Removal in Machine Learning Leads to Leakage

Machine learning (ML) approaches to data analysis are now widely adopted...

Please sign up or login with your details

Forgot password? Click here to reset