Data Science Weekly Newsletter

Issue

271

January 31, 2019

‍

Editor's Picks

‍

Why are Machine Learning Projects so Hard to Manage?
I’ve watched lots of companies attempt to deploy machine learning — some succeed wildly and some fail spectacularly. One constant is that machine learning teams have a hard time setting goals and setting expectations. Why is this?...

Predicting a Startup Valuation with Data Science
The following is a condensed and slightly modified version of a Radicle working paper on the startup economy in which we explore post-money valuations by venture capital stage classifications. We find that valuations have interesting distributional properties and then go on to describe a statistical model for estimating an undisclosed valuation with considerable ease...

Fuelled by Data: How to Pace the London Marathon
As a self-confessed running addict, this was an opportunity to dive into some juicy running data. And which race to analyse? Only the best race in the world would do: the London Marathon...

‍

A Message From This Week's Sponsor

‍

The sins of recruitment

These days you’re likely to encounter bad dating behaviours in the hunt for your dream job as you are for your dream date. In fact a shocking 90% of job hunters have claimed to experience one of the notorious ‘recruiting sins’.

‍

Data Science Articles & Videos

‍

Towards reconstructing intelligible speech from the human auditory cortex
Thanks to fMRI scanning, we’ve known for decades that when people speak, or hear others, it activates specific parts of their brain. However, it’s proved hugely challenging to translate thoughts into words. A team from Columbia University has developed a system that combines deep learning with a speech synthesizer to do just that...

Making face recognition less biased doesn’t make it less scary
Three new studies propose ways to make algorithms better at identifying people in different demographic groups. But without regulation, that won’t curb the technology’s potential for abuse...

Go-Explore: A New Type of Algorithm for Hard-exploration Problems
[Jeff Clune Talk]
This presentation took place at the Deep Learning Summit, San Francisco 24th January 2019. Jeff Clune is a Senior Research Scientist & Founding Member at Uber AI Labs. He focuses on robotics, reinforcement learning and training neural networks either via deep learning or evolutionary algorithms...

Modders are using AI to upscale pre-rendered PS1 backgrounds with phenomenal results
This may be the closest we ever get to "HD" remasters of early 3D games. Here's what ESRGAN is, and why it's so exciting...

Ganbreeder: Create beautiful, wild and weird images
A new Ganbreeder is live. Higher resolution, built in 1024px upscaling, user accounts, plenty of UI improvements and still free to use! Thanks everyone for the feedback. Enjoy :)...

Obstacle Tower Environment
The Obstacle Tower is a procedurally generated environment consisting of multiple floors to be solved by a learning agent. It is designed to test learning agents abilities in computer vision, locomotion skills, high-level planning, and generalization. It combines platforming-style gameplay with puzzles and planning problems, and critically, increases in difficult as the agent progresses...

Transformer-XL: Unleashing the Potential of Attention Models
Introducing Transformer-XL, a novel architecture that enables natural language understanding beyond a fixed-length context...

On the Effectiveness of Interval Bound Propagation for Training Verifiably Robust Models
Through a comprehensive analysis, we show how a careful implementation of a simple bounding technique, interval bound propagation (IBP), can be exploited to train verifiably robust neural networks that beat the state-of-the-art in verified accuracy...

‍

Jobs

‍

Quantitative Behavioral Scientist - BetterUp - San Francisco, remote ok

BetterUp Labs is currently seeking an innovative, early-career quantitative behavioral scientist who is passionate about advancing our understanding of the inner life and whole person performance of professionals around the globe. You will help design and implement original research to learn more about what makes us tick when we’re at work. You’ll need to draw on your budding experience as an experimental social scientist, inferential statistician, computational scientist, and lover of all things Data to uncover the truly groundbreaking answers to these questions...

Want to post a job here? Email us for details >> team@datascienceweekly.org

‍

Training & Resources

‍

How to Subclass The nn.Module Class in PyTorch
Learn how to construct a Custom PyTorch Model by creating your own custom PyTorch module by subclassing the PyTorch nn.Module class, via a screencast video and full tutorial transcript...

What are Symbolic and Imperative APIs in TensorFlow 2.0?
In this article, I’ll explain the tradeoffs between two styles you can use to create your neural networks...

‘Diversity in Faces’ Dataset to
Advance Study of Fairness in Facial Recognition Systems
Today, IBM Research is releasing a new large and diverse dataset called Diversity in Faces (DiF) to advance the study of fairness and accuracy in facial recognition technology. The first of its kind available to the global research community, DiF provides a dataset of annotations of 1 million human facial images...

‍

Books

‍

The Book of R: A First Course in Programming and Statistics
"The Book of R is a comprehensive, beginner-friendly guide to R, the world’s most popular programming language for statistical analysis. Even if you have no programming experience and little more than a grounding in the basics of mathematics, you’ll find everything you need to begin using R effectively for statistical analysis"...

For a detailed list of books covering Data Science, Machine Learning, AI and associated programming languages check out our resources page

‍