https://peer.adalta.social/w/hyzCGtuEbwUouFsyhLDExE [](https://p4u.xyz/ID_TQHMTARM/1)Die fundamentale Schwachstelle liegt in den Trainingsdaten, nicht in den Algorithmen#technology #bot #war #python #data
https://peer.adalta.social/w/sqdb9P4gkBy7sn9EFkMUsc [](https://p4u.xyz/ID_O-Q12NQD/1)Die fundamentale Schwachstelle liegt in den Trainingsdaten, nicht in den Algorithmen#technology #bot #war #python #data
https://peer.adalta.social/w/9Meq8szC5mFBRvZQNCCA98 [](https://p4u.xyz/ID_WHODM-ET/1)Comment les références de données mutables sabotent les projets de ML#technology #bot #war #python #data
https://peer.adalta.social/w/hQJwyfuJw6ah5wkq8DjG12 [](https://p4u.xyz/ID_M76JK3MR/1)L'instabilité des résultats provient majoritairement de problèmes de données, non de modèles.#technology #war #python #data #train
https://peer.adalta.social/w/9XJXtHLzvZvB4igWBuzQYz [](https://p4u.xyz/ID_M76JK3MR/1)Mutable Data References and the Hidden Instability of ML Pipelines#technology #war #python #data #train
Most ML issues are not model problems. They are data problems.I retrained the same churn model twice.Same code. Same path to the data.Different result.Why? Because of mutable data references. I wrote a small Data Lake vs Data Lakehouse demo showing why versioned data makes ML debugging reproducible: https://tinyurl.com/lake-vs-lakehouse-medium Friend-Link: https://medium.com/towards-artificial-intelligence/from-data-lake-to-data-lakehouse-why-ai-changes-the-rules-for-data-platforms-c78feab48e1c?sk=405811cbc10baa4622bcfcad90736ed4#ai #machinelearning #data #lakehouse #warehouse #python #datalake #technology #regression
