Years ago, when we talked about open source code, we were talking about something that was really only relevant to classic computer geeks – the ones running their computers on Unix and swapping bits of code on message boards. Times have changed, though, and now open source code is an integral part of how most businesses and just about anyone in the software world does their work.Read Full Story
I just pushed some of my structured learning code to github and hope that some people might find it useful. Before describing my code here, I wanted to give a basic intro into structured prediction. I hope I can at least convey some intuition for this vast research area.Read Full Story
Summary: I learn best with toy code that I can play with. This tutorial teaches gradient descent via a very simple toy example, a short python implementation.
Followup Post: I intend to write a followup post to this one adding popular features leveraged by state-of-the-art approaches (likely Dropout, DropConnect, and Momentum). I’ll tweet it out when it’s complete @iamtrask.
We review the math and code needed to fit a Gaussian Process (GP) regressor to data. We conclude with a demo of a popular application, fast function minimization through GP-guided search. The gif below illustrates this approach in action — the red points are samples from the hidden red curve. Using these samples, we attempt to leverage GPs to find the curve’s minimum as fast as possible.Read Full Story
Last week I shared a thought on Twitter:
When you’ve written the same code 3 times, write a functionWhen you’ve given the same in-person advice 3 times, write a blog post— David Robinson (@drob) November 9, 2017
Ironically, this tweet hints at a piece of advice I’ve given at least 3 dozen times, but haven’t yet written a post about.
Code Reviews, Dogfooding, Deobfuscation, and Differential Privacy
How to Do a Code Review — Google’s guidelines. Encourage developers to solve the problem they know needs to be solved now, not the problem that the developer speculates might need to be solved in the future.
Code Monster from Crunchzilla is now open source, free to use and modify. Code Monster is a tutorial that has been used by hundreds of thousands of children around the world to learn a little about programming. It’s a series of short lessons where each lesson involves reading and modifying a small amount of code.Read Full Story
There were many great talks again at Oracle Open World & Code One 2018! See the OOW’18 and Code One’18 Sessions Catalogs to see everything there.
For easy access and downloads, I’m posting my Oracle’s Machine Learning talks, abstracts, downloads and “live” recording below.
Neo4j 3.0 introduced the concept of user defined procedures: code written in Java (or any JVM language) that is deployed to the database and callable from Cypher. User defined procedures are an alternative to unmanaged extensions, with the key difference that user defined procedures are callable from Cypher (instead of extending the http REST endpoints).Read Full Story
Eleven months ago on a long train ride home, I wrote the first lines of code for a small platforming game. Little did I know that this prototype was the start of something much more than a just game — it was a dream that would become shared within an amazing team, and it was the greatest step in a personal journey that had begun over eight years ago.Read Full Story
Summary: I learn best with toy code that I can play with. This tutorial teaches DeepMind’s Neural Stack machine via a very simple toy example, a short python implementation. I will also explain my thought process along the way for reading and implementing research papers from scratch, which I hope you will find useful.Read Full Story
Summary: I learn best with toy code that I can play with. This tutorial teaches Recurrent Neural Networks via a very simple toy example, a short python implementation. Chinese Translation Korean Translation
I’ll tweet out (Part 2: LSTM) when it’s complete at @iamtrask.
@drsimonj here to share my code for using Welch’s t-test to compare group means using summary statistics.
I’ve just started working with A/B tests that use big data. Where once I’d whimsically run t.test(), now my data won’t fit into memory!
I’m sharing my solution here in the hope that it might help others.
Welcome to part 2 of the JackBlockChain, where I write some code to introduce the ability for different nodes to communicate.
Initially my goal was to write about nodes syncing up and talking with each other, along with mining and broadcasting their winning blocks to other nodes.
This notebook and code are available on Github.
This notebook illustrates a Tensorflow implementation of the paper “A Neural Algorithm of Artistic Style” which is used to transfer the art style of one picture to another picture’s contents.
If you like to run this notebook, you will need to install TensorFlow, Scipy and Numpy.
You will need to download the VGG-19 model.
Data science interviews are the worst because data science is interdisciplinary: code for “you have to know everything about all the disciplines.” Depending on the company and the team, your interview might look like a software developer’s interview, or it might look a like a statistician’s interview, and the bad news is that virtually none of the material overlaps.Read Full Story
We give a simple class for marking the time at different points in a code block and then printing out the time gaps between adjacent marked points. This is useful for identifying slow spots in code.Read Full Story
Get started with reinforcement learning in less that 200 lines of code with
Keras (Theano or Tensorflow, it’s your choice).
So you are a (Supervised) Machine Learning practitioner that was also sold the
hype of making your labels weaker and to the
possibility of getting neural networks to play your favorite games.
Docker Container is stand-alone software that contains both application code and its dependencies, which can run in any platform smoothly. Docker uses Linux namespaces and Cgroups to isolate different containers. The motto of Docker is to build once anywhere and run anywhere. You can use Docker to create an image, run it as a container, and ship it anywhere.Read Full Story
ICML, ICLR, and NeurIPS are all considering or experimenting with code and data submission as a part of the reviewer or publication process with the hypothesis that it aids reproducibility of results. Reproducibility has been a rising concern with discussions in paper, workshop, and invited talk.
The fundamental driver is of course lack of reproducibility.