External Course

The psychological roots of intellectual humility: The role of intelligence and cognitive flexibility

DSA ADS Course - 2021

Psychological roots of intellectual humility: The role of intelligence and cognitive flexibility

Human Decision Making, Intelligence, Cognitive Flexibility, Confirmation Bias, Intellectual Humility

Cognitive flexibility and having a diverse toolbox of conceptual frameworks is critical to high performance data science. Humility is important to admitting error and self correcting in near real-time situations.

Cognitive flexibility and humility is magic sauce to avoiding and mitigating confirmation bias and optimal decision-making.

Causal Inference and the Data-fusion Problem

DSA ADS Course - 2021

Causal Inference, Data-fusion, Causal Analysis, Causal Data Science, Counterfactuals, Selection Bias

Discuss concepts and techniques of different approaches to causal analysis, the curse of big data, data fusion and bias challenges. Discuss biases such as: confounding, sampling selection, and cross-population biases, along with a general, potential mitigation nonparametric framework for handling biases.

Discuss appropriate use of counterfactuals in causal analysis and risk of reasonable inference vs. unreasonable inference.

Counterfactual Data-Fusion for Online Reinforcement Learners

DSA ADS Course - 2021

Causal Reinforcement Learning, Counterfactuals, Counterfactual Data-Fusion, Online Reinforcement Learners

Discuss counterfactuals, risks of data fusion, causal reinforcement learning and fusion of observational and experimental data.

Counterfactual Data-Fusion for Online Reinforcement Learners - June, 2017


A Critical View of the Structural Causal Model

DSA ADS Course - 2021

Causal Inference, Causality, Structural Causal Model, SCM

Discuss causality vs. correlation - causality vs. statistics in machine learning. Review techniques for attempting to find true causality. 

In statistics, correlation or dependence is any statistical relationship, whether causal or not, between two random variables or bivariate data. In the broadest sense correlation is any statistical association, though it commonly refers to the degree to which a pair of variables are linearly related.

Models Only Say What They’re Told to Say

July, 2021


This short note is a reminder that all models only say what they are told to say. There can therefore be no discovery by models—discoveries go into models—and thus models should only be used in their predictive form, and only trusted when they have demonstrated independent success.

See also: Proper Use and Misuse of Modeling - DSA ADS Course 2022

DSA ADS Course - 2022