Machine Learning on Go Code

Explore machine learning applications for Go programming in this GopherCon 2018 talk by Francesc Campoy Flores. Discover how ML techniques can enhance Go development, from predicting characters to identifying potential bugs. Learn about embeddings for identifiers and source code, recurrent neural networks for code completion, and future research directions. Gain insights into the advantages and limitations of applying ML to code, with minimal mathematical complexity. Understand how these techniques could impact developer workflows and improve code quality. Delve into topics like data processing, static analysis, neural networks, and automated code review. Consider the implications of ML on software development and how it may empower rather than replace developers.

Intro
Photoshop
Windows NT 31
Apache OpenOffice
Windows XP
I 150 million lines of code
That sounds dangerous
Googles code
Tools
Better Tools
Who am I
YouTube channel
Agenda
What is Machine Learning
Related Fields
What does this require
First challenge
Data sets
Processing
Language
Functions
Declaration
Data Analysis
Source Code
Character by Character
Token
Static Analysis
Neural Networks
ML Encode
Recurrent Neural Networks
Character Recurring Neural Network
Before Training
Go
Generating Natural Language
VAR Misuse
Google Slides
Imagine
Recommendation
Automated Code Review
Bug Prediction
Will Developers Be Replaced
Developers Will Be Empowered
Learn More