Tutorial

Getting started > Dependencies > Building > Installing > [Tutorial]

---

Tutorials on VowpalWabbit.org

• Getting started
  • Introduction to VW concepts
  • Uses command line interface
• Contextual Bandits
  • Introduction to contextual bandit algorithms
  • Uses Python
• Simulating a news personalization scenario using Contextual Bandits
  • Setup a simulation scenario using a contextual bandit algoritm
  • Uses Python

More Tutorials

General

• High level introduction of Vowpal Wabbit (input format, useful options and more) through the lens of logistic regression, by Philippe Adjiman.

The following tutorials generally cover features added in each release, but may be slightly outdated due to their age.

Version 8.5 Tutorials - NIPS 2017

• Sparse model, baseline, optimized exploration
• Cost Sensitive Active Learning
• Java interface
• Decision Service JSON Ingestion

Version 8.3 Tutorials - NIPS 2016

• Contextual bandits
• Logarithmic time classification
• Joint Prediction.

Version 8.1 Tutorials - NIPS 2015

• Learning to search results + active learning + C# library + decision service

Version 7.8 Tutorials

• Intro + reductions + log_multi + exploration library + AzureML
• Polynomial learning
• LRQ + Hogwild mode
• Online kernel SVM
• Learning2Search + python
• Entity Relation and Dependency Parsing

Version 7.4 Tutorials

• The Learning Reduction and Searn systems
• Several improvements from Zhen including holdout, bootstrap, early termination, top k
• Normalized Gradient Descent from Paul.
• A new Active Learning interface from Nikos.

Version 7.0 Tutorial

• Version 7.0 tutorial
• It covers the basics and most common options, how to use VW and the data format for different types of problems, such as Binary Classification, Regression, Multiclass Classification, Cost-Sensitive Multiclass Classification, "Offline" Contextual Bandit and Sequence Predictions. Many more advanced options in terms of flags and the data format are not covered. You can refer to previous tutorials for these more advanced details.

Version 6.1 Tutorials

• This tutorial and other items below cover some topics that weren't covered in version 7 as they haven't changed in that version.
• Version 6.1 tutorial introduction
• Description and use of L-BFGS.
• Cluster parallel learning.
• Active Learning (v5.0 presentation, but little changed)
• Latent Dirichlet Allocation (v5.0 presentation, little changed) See also Latent Dirichlet Allocation
• Vowpal Wabbit tutorial for the Uninitiated by Rob Zinkov

Version 5.1 Tutorials

• Version 5.1 tutorial
• Video

Version 5.0 Tutorials

• Version 5.0 Videolecture
• Version 5.0 tutorial
• The importance weight invariant update rule. (covered in 6.1 intro)