Observability and DKIW model


DKIW (Data, Knowledge, Information and Wisdom) model, identifies the path to extract meaningful information from raw data and convert it into rules for decisions. For system observability, this is quite interesting. There is normal levels of maturity in this systems. No recollect logs/data at all Recollect logs but not centralized (Data) Recollect logs centralized, but without alarming (Information) Creation of metrics for health systems (Wisdom) -> lots of times with alarms References Data Demystified — DIKW model - https://towardsdatascience.…
Read more ⟶

Descendint Through a Crowded Valley - Benchmarkiing Deep learning optimizers


From https://www.youtube.com/watch?v=DiNzQP7kK-s A paper that does some benchmarking what optimizers to use. One thing I liked is the list of all the optimizers and schedulers there are. Conclusion ::: anki 34307583-9dce-48e7-b0f4-e4e59059368e Which is the best optimizer for DL? Adam or SGD with momentum #Flashcard :::…
Read more ⟶

Crucial Conversation


Chapter 1 :::: anki 6365a6b2-7296-482f-af37-1164cb207054 What is a crucial conversation, one that has: high stakes, emotions and opposing opinions #Flashcard ::: Chapter 2 - Power of dialog ::: anki 9cd0c219-28f1-448d-bea3-ad6b27051a24 Fool’s choice paradox you have only two options, tell the truth and make an enemy or suffer in silence #Flashcard ::: 👆 is a lie! Fill the pool of shared meaning People who are skilled at dialoge, do their best to make it safe for everyone to add meaning to the pool of shared meaning The pool of shared meaning is the birtplace of Sinergy…
Read more ⟶