Grigori Fursin
2020-08-28
<p><strong>Invited talk at FastPath 2020 (International Workshop on Performance Analysis of Machine Learning Systems) co-located with ISPASS 2020.</strong></p>
<p><strong>Program: </strong><a href="https://fastpath2020.github.io/Program">fastpath2020.github.io/Program</a></p>
<p><strong>Abstract:</strong></p>
<p>10 years ago we released our novel ML-based MILEPOST compiler with all the related experimental data at <a href="http://ctuning.org">cTuning.org</a>. Unfortunately, this research quickly stalled after we struggled to reproduce performance results and predictive models shared by volunteers across rapidly changing systems.</p>
<p>In this talk, I will describe my 10-year effort to solve numerous reproducibility issues in ML&systems research. I will share my experience reproducing 150+ systems and ML papers during artifact evaluation at ASPLOS, MLSys, CGO, PPoPP and Supercomputing. This tedious experience motivated me to develop the <a href="http://cknowledge.org">cKnowledge.org</a> framework and the open <a href="https://cKnowledge.io">cKnowledge.io</a> portal to bring DevOps principles to our research. I will also present cKnowledge solutions - a new way to package and share research artifacts and results with common Python APIs, CLI actions, portable workflows and JSON meta descriptions. Such solutions can be used to automatically build, benchmark and validate ML&system experiments across continuously evolving platforms.</p>
<p>I will conclude with <a href="https://cKnowledge.org/partners">several practical use-cases</a> of our technology in collaboration with Arm, IBM, General Motors, the Raspberry Pi foundation and MLPerf. Our long-term goal is to help researchers share their new ML techniques as production-ready packages along with published papers and participate in collaborative and reproducible benchmarking, co-design and comparison of efficient ML/software/hardware stacks.</p>
<p> </p>
https://doi.org/10.5281/zenodo.4005588
oai:zenodo.org:4005588
eng
Zenodo
https://zenodo.org/communities/ck
https://doi.org/10.5281/zenodo.4005587
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
reproducibility
reusability
portability
workflow
automation
machine learning
systems
benchmarking
artifact evaluation
fair principles
open science
Enabling reproducible ML&systems research: the good, the bad, and the ugly
info:eu-repo/semantics/lecture