[Colloquium] Ray Sinurat MS Presentation/Feb 6, 2024

[meganwoodward at uchicago.edu]

This is an announcement of Ray Sinurat's MS Presentation
===============================================
Candidate: Ray Sinurat

Date: Tuesday, February 06, 2024

Time: 10 am CST

Location: JCL 390

Title: Towards Continually Learning Application Performance Models

Abstract: Machine learning-based performance models are increasingly being
used to build critical job scheduling and application optimization decisions.
Traditionally, these models assume that data distribution does not change as
more samples are collected over time. However, owing to the complexity and
heterogeneity of production HPC systems, they are susceptible to hardware
degradation, replacement, and/or software patches, which can lead to drift in the
data distribution that can adversely affect the performance models. To this end,
we develop continually learning performance models that account for the distribution
drift, alleviate catastrophic forgetting, and improve generalizability. Our best model
was able to retain accuracy, regardless of having to learn the new distribution of data
inflicted by system changes, while demonstrating a 2x improvement in the prediction
accuracy of the whole data sequence in comparison to the naive approach.


Advisors: Haryadi Gunawi

Committee Members: Haryadi Gunawi, Sandeep Madireddy, Kexin Pei