Powered by Jupyter Book

Back to top

Sparsh A.

Index

Creative Commons License – This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International.

Contents

Off-Policy Evaluation

Concepts

Counterfactual policy evaluation
Evaluating non-stationary policies
Online vs Offline Evaluation

Datasets

Open Bandit Dataset

Tools

OBP Library

Models

Direct method
Inverse Propensity Weighting
Doubly Robust

Tutorials

Offline Policy Evaluation with VW Command Line
OBP Library Workshop Tutorials
Evaluating a New Fraud Policy with DM, IPW, and DR Methods
Evaluating Standard Off-Policy Estimators with Small Sample Open-Bandit Dataset
Evaluation of Multiple Off-Policy Estimators on Synthetic Dataset
Off-Policy Learning in Two-stage Recommender Systems
Batch Learning from Bandit Feedback
Evaluating the Robustness of Off-Policy Evaluation
Adaptive Estimator Selection for Off-Policy Evaluation
Optimal Off-Policy Evaluation from Multiple Logging Policies

Lecture (PDF)
Book (PDF)

Notebook Launcher

Choose public or private cloud service for "Launch" button.

Select a server

Public
Private

Launch Notebook