Spice up your website with Machine Learning! Evelina Gabasova - PowerPoint PPT Presentation

Spice up your website with Machine Learning! Evelina Gabasova @evelgab

F# Snippets

F# Snippets fssnip.net

Searching through F# snippets over 1600 snippets over 1100 different tags

Searching through F# snippets

Do we need a custom system?

Great opportunity to create a custom machine learning system!

Nguyen A et al.: Deep Neural Networks are Easily Fooled: High Confidence Predictions for Unrecognizable Images. 2015.

Using machine learning in production dependence on training data inputs

User-generated inputs

PART I Finding related snippets If you liked this F# code, you'll also like ...

Simple information retrieval common terms

Bag of words ignore order of words separate text and code

Term frequency Snippet 1 Snippet 2 Term Frequency Term Frequency async 3 async 0 x 15 x 15 The 2 The 2 code 1 code 1 ... ...

Inverse document frequency Relative importance of terms number of snippets idf (term) = log number of snippets with term

Vector representation: TF-IDF Term frequency - inverse document frequency tfidf (term, snippet) = tf (term, snippet) × idf (term)

Demo

Vector representation of snippets Snippet x List Array ... snippet1 0 0.17 0 ... snippet2 0 0.04 0.001 ... snippet3 0.23 0.005 0.31 ... snippet4 0 0 0 ... ...

Vector representation of snippets

PART II Suggesting tags

Suggesting tags

Making sense of user-generated tags async, #async, async mailprocessor, async paraller, Async sequences, asyncseq, asynchronous, Asynchronous Processing, Asynchronous Programming, asynchronous sequence, asynchronous workflows

Edit distance regex vs. regexp sports vs. ports pi vs. API

Machine learning From snippets to tags

Associations string and parser async and MailboxProcessor sequence and exception

Naive Bayes Why do you call me naive?

Why naive? string and parser async and MailboxProcessor sequence and exception

Building a predictor

Building a predictor

Building a predictor

Tag probabilities Bayes theorem p ( A ∣ B ) = p ( B ∣ A ) p ( A ) p ( B )

Tag probabilities Bayes theorem p (tag ∣ snippet) ∝ p (tag) p (snippet ∣ tag)

Tag probabilities Bayes theorem p (tag ∣ snippet) ∝ p (tag) p (term ∣ tag) ∏ term

1. Prior probabilities p (tag) ≈ Number of snippets with the tag Number of snippets

2. Tag likelihood How frequent is the term among snippets that have the tag ? p (term ∣ tag) = Number of snippets with the term and tag Number of snippets with the tag

Naive Bayes prediction p (tag ∣ snippet) ∝ p (tag) p (term ∣ tag) ∏ term p (tag ∣ snippet) > 1? p (¬tag ∣ snippet)

The theory is always nicer What if there is no snippet tagged async that contains List?

Demo

Do you really need a custom system? Domain representation What are important features Machine learning is fun!

Learning more F# snippets fssnip.net F# snippets on GitHub github.com/fssnippets The F# Foundation www.fsharp.org FsLab Package www.fslab.org Introduction to information retrieval informationretrieval.org

Workshop Polyglot Data Science: The Force Awakens Friday, April 1 Data science, F#, R, D3.js ... and Star Wars!

Thank you! @evelgab github.com/evelinag evelinag.com