Pinnedfabio casatiThe Accuracy Paradox in Machine LearningAnd why the implications of a correct ML model assessment will redefine what we understand with “learning” in ML.Jan 6, 20224Jan 6, 20224
fabio casatiAnnoying frequent perspectives on responsible AI(a subject nobody cared about until basically yesterday, as shown below, but that now includes many experts)May 30, 2023May 30, 2023
fabio casatiHow “Good” is Your ML Model?A lot of R&D work goes today in building better ML models and pipelines, both in academia and in industry. Metrics, from F1 to BLEU score…Oct 11, 2022Oct 11, 2022
fabio casatiTrain-Test Data SplitThere is an enormous amount of confusion on how to train-test split the data, from wrong reasons for splitting to wrong assumptions. Many…Oct 4, 2022Oct 4, 2022
fabio casatiThe Most Common ML Quality Engineering MistakesEngineering quality in AI systems today — or even understanding what quality means — is still an art, and an art which few companies…Sep 30, 2022Sep 30, 2022
fabio casatiEnterprise AI Must be Boring — or it will failOn the properties of “good” enterprise AI systemsAug 26, 20221Aug 26, 20221
fabio casatiAn Opinionated Tutorial on AI and Societal Fairness — Part 3: on Fairness and BiasI suggest reading Part 1 and Part 2 before diving into fairness and bias considerations.Jun 16, 2022Jun 16, 2022
fabio casatiAn Opinionated Tutorial on AI and Societal Fairness — Part 2: why do we have AI BiasI suggest reading Part 1 first, where we go through examples that show what started the conversation on AI bias and what are the different…Jun 16, 2022Jun 16, 2022