Jean Czerlinski Whitmore Ortega – Medium

Jean Czerlinski Whitmore Ortega

Jean Czerlinski Whitmore Ortega

Constraining the gaming that scales

When one person’s gaming scales into crisis

May 1

Constraining the gaming that scales

May 1

Jean Czerlinski Whitmore Ortega

Why incentive advice diverges: Is failure likely but acceptable?

Unattainable goals with unacceptable consequences spur workers to game incentives rather than give up

Mar 11

Why incentive advice diverges: Is failure likely but acceptable?

Mar 11

Jean Czerlinski Whitmore Ortega

When models get gamed: changing incentives to fool a model

Spam bots, credit card fraud, and the history of search engines

Jul 31, 2023

When models get gamed: changing incentives to fool a model

Jul 31, 2023

Jean Czerlinski Whitmore Ortega

Two ways human bias gets into — and out of — ChatGPT

The role of humans in training large language models

May 30, 2023

Two ways human bias gets into — and out of — ChatGPT

May 30, 2023

Jean Czerlinski Whitmore Ortega

Strategic classification in the dark (a review)

When public algorithms are more robust to manipulation than secret algorithms

Feb 9, 2023

playing chess in a dark room

Feb 9, 2023

Jean Czerlinski Whitmore Ortega

Spotlight on the bias-variance trade-off

Why more data is not always better

Sep 6, 2022

U-shaped test error curve.

Sep 6, 2022

Jean Czerlinski Whitmore Ortega

How millions of parameters can avoid overfitting

Both linear regression and deep learning can leverage a massive number of mis-specified features

Jan 3, 2022

How millions of parameters can avoid overfitting

Jan 3, 2022

Jean Czerlinski Whitmore Ortega

Social chess: stakeholders with impactful feedback loops

How machine learning models are embedded in a web of strategic moves

Nov 8, 2021

Social chess: stakeholders with impactful feedback loops

Nov 8, 2021

Jean Czerlinski Whitmore Ortega

When model builders can — and can’t — rest on their laurels

Jennifer Lopez’ dress was exogenous change while image spam was endogenous change

Nov 2, 2021

When model builders can — and can’t — rest on their laurels

Nov 2, 2021

Jean Czerlinski Whitmore Ortega

When the irrelevant is made relevant

What the !@# does it mean to “break” a model?

Oct 23, 2021

When the irrelevant is made relevant

Oct 23, 2021

Jean Czerlinski Whitmore Ortega

Jean Czerlinski Whitmore Ortega

Ex-Google engineer modeling things and celebrating non-things: machine learning, incentives, behavior, ethics, physics.

Following

Help
Status
About
Careers
Press
Blog
Privacy
Terms
Text to speech
Teams