PinnedValerie CareyinTowards Data ScienceWhat (Not) To Say When Your Client Questions Your ResultsUse these pivotal moments to build trust and understanding·12 min read·Jul 19, 2022--4--4
Valerie CareyinTowards Data ScienceNo Label Left Behind: Alternative Encodings for Hierarchical CategoricalsSeeking a system that works for current and future codes·15 min read·May 17, 2024----
Valerie CareyinTowards Data ScienceExploring Hierarchical Blending in Target EncodingWhen can code hierarchies improve target encoding for high-cardinality categorical features?·12 min read·Apr 18, 2024--2--2
Valerie CareyinTowards Data ScienceSHAP vs. ALE for Feature Interactions: Understanding Conflicting ResultsModel Explainers Don’t Produce Explanations·10 min read·Oct 2, 2023--1--1
Valerie CareyinFanfareBook Discussion| AI 2041: Ten Visions for Our FutureExceptional young men, lovelorn young women, and bots·13 min read·May 14, 2022--1--1
Valerie CareyinTowards Data ScienceAI Integrity: Leadership Lessons from Other IndustriesDo other fields make mistakes better?11 min read·Feb 4, 2022--1--1
Valerie CareyinTowards Data ScienceAI Integrity: Planning Ahead to Do The Right ThingHow to prepare for inevitable mistakes12 min read·Nov 30, 2021--1--1
Valerie CareyinTowards Data ScienceFeature Choice and Fairness: Less May be MoreThoughtful predictor selection is essential for model fairness10 min read·Mar 15, 2021----
Valerie CareyinTowards Data ScienceHow to Fix Feature BiasChoosing a strategy requires testing and tradeoffs12 min read·Feb 8, 2021--1--1