There is a newer version of this record available.

Presentation Open Access

Enabling reproducible ML&systems research: the good, the bad, and the ugly

Grigori Fursin

Invited talk at FastPath 2020 (International Workshop on Performance Analysis of Machine Learning Systems) co-located with ISPASS 2020.

Program: fastpath2020.github.io/Program

Abstract:

10 years ago we released our novel ML-based MILEPOST compiler with all the related experimental data at cTuning.org. Unfortunately, this research quickly stalled after we struggled to reproduce performance results and predictive models shared by volunteers across rapidly changing systems.

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 cKnowledge.org framework and the open cKnowledge.io 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.

I will conclude with several practical use-cases 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.

 

Files (4.0 MB)
Name Size
presentation-fastpath.pdf
md5:a2b7971b4143b7516fabdb51000ce363
4.0 MB Download
  • Grigori Fursin (2020). The Collective Knowledge project: making ML models more portable and reproducible with open APIs, reusable best practices and MLOps. arXiv:2006.07161

4,890
3,450
views
downloads
All versions This version
Views 4,89037
Downloads 3,45020
Data volume 14.8 GB81.0 MB
Unique views 4,18633
Unique downloads 3,16617

Share

Cite as