reproducibility – Andrew Fairless, Ph.D.

What I Read: Machine Learning Metadata

By Andrew Fairless on October 24, 2022September 9, 2022

https://www.kdnuggets.com/2022/08/machine-learning-metadata-store.html Machine Learning Metadata StoreBy Akruti Acharya, Technical Content WriterAugust 31, 2022 “Storing the metadata would help in retraining the same model and get the same results…. Hence, the needContinue readingWhat I Read: Machine Learning Metadata

What I Read: Modern Stack for ML Infrastructure

By Andrew Fairless on June 13, 2022May 23, 2022

https://youtu.be/gZnhSHvhuFQThe Modern Stack for ML Infrastructure | OuterboundsMay 2, 2022Data Council “Metaflow was originally developed at Netflix to provide a user-friendly platform for a wide range of ML use cases…Continue readingWhat I Read: Modern Stack for ML Infrastructure

What I Read: Maps of Model Space, Stan

By Andrew Fairless on January 5, 2022December 11, 2021

https://statmodeling.stat.columbia.edu/2021/11/19/drawing-maps-of-model-space-with-modular-stan/ Drawing Maps of Model Space with Modular StanPosted on November 19, 2021 9:50 AM by AndrewThis post is by Ryan Bernstein. “What if instead of describing one model atContinue readingWhat I Read: Maps of Model Space, Stan

What I Read: AntiPatterns, MLOps

By Andrew Fairless on January 4, 2022December 11, 2021

https://arxiv.org/abs/2107.00079 Using AntiPatterns to avoid MLOps MistakesNikhil Muralidhar, Sathappah Muthiah, Patrick Butler, Manish Jain, Yu Yu, Katy Burne, Weipeng Li, David Jones, Prakash Arunachalam, Hays ‘Skip’ McCormick, Naren Ramakrishnan “HereContinue readingWhat I Read: AntiPatterns, MLOps

What I Read: Lessons on ML Platforms

By Andrew Fairless on December 20, 2021November 16, 2021

https://towardsdatascience.com/lessons-on-ml-platforms-from-netflix-doordash-spotify-and-more-f455400115c7 Lessons on ML Platforms — from Netflix, DoorDash, Spotify, and moreErnest ChanMay 12 “Many companies who do ML well have invested in ML platforms to enable data scientists toContinue readingWhat I Read: Lessons on ML Platforms

What I Read: The Uselessness of Useful Knowledge

By Andrew Fairless on November 2, 2021October 24, 2021

https://www.quantamagazine.org/science-has-entered-a-new-era-of-alchemy-good-20211020/ The Uselessness of Useful KnowledgeRobbert DijkgraafOctober 20, 2021Today’s powerful but little-understood artificial intelligence breakthroughs echo past examples of unexpected scientific progress.“…the current state of AI research is nothing newContinue readingWhat I Read: The Uselessness of Useful Knowledge

What I Read: I Like Notebooks

By Andrew Fairless on August 17, 2021July 22, 2021

https://youtu.be/9Q6sLbz37gk I Like NotebooksOct 28, 2020Jeremy Howard “I like using Jupyter Notebooks… Particularly when combined with nbdev… In this video, I explain why, and explain why I have a differentContinue readingWhat I Read: I Like Notebooks

What I Read: null hypothesis significance testing falls apart when considering replications

By Andrew Fairless on July 8, 2021June 22, 2021

https://statmodeling.stat.columbia.edu/2021/06/17/guttman-points-out-another-problem-with-null-hypothesis-significance-testing-it-falls-apart-when-considering-replications/ Guttman points out another problem with null hypothesis significance testing: It falls apart when considering replications.by Andrew Gelman17 June 2021, 9:22 am “… another problem with null hypothesis significanceContinue readingWhat I Read: null hypothesis significance testing falls apart when considering replications

What I Read: technical debt, ML pipelines

By Andrew Fairless on March 17, 2021March 7, 2021

https://towardsdatascience.com/avoiding-technical-debt-with-ml-pipelines-3e5b6e0c1c93?gi=8f8c9d302be Avoiding technical debt with ML pipelinesStrike a balance between rapid results and high-quality coding.Hamza Tahir “ML practitioners in many organizations are heavily incentivized to make quick-wins to produce earlyContinue readingWhat I Read: technical debt, ML pipelines

Tag: reproducibility