[in case you missed it] Data Science Weekly - Issue 310
Issue #310 Oct 31 2019
Editor Picks
Rising Seas Are Going to Drown Way More Cities Than We’d Thought: Study
In a new study published by the journal Nature Communications, scientists affiliated with the organization Climate Central and Princeton University detail previous methodological problems, then use artificial intelligence to determine — and correct for — the previous literature’s error rate. Their research yields some eye-popping (or stomach turning) updates to our conventional understanding of what the next century has in store for our coastlines...
A neural net solves the three-body problem 100 million times faster
Machine learning provides an entirely new way to tackle one of the classic problems of applied mathematics...
DeepMind’s AI has now outcompeted nearly all human players at StarCraft II
AlphaStar cooperated with itself to learn new strategies for conquering the popular galactic warfare game...
A Message from this week's Sponsor:
Data scientists are in demand on Vettery
Vettery is an online hiring marketplace that's changing the way people hire and get hired. Ready for a bold career move? Make a free profile, name your salary, and connect with hiring managers from top employers today.
Data Science Articles & Videos
Daemon Spawn, AGI Takeover, Deepfake Deluge, Bias Crisis -
How Scared Should You Be?
I promised last week to share some common reasons for AI project failures. But first, let’s start with some of the least common reasons...
Model Parameters and Hyperparameters in Machine Learning —
What is the difference?
What makes the difference between a good and a bad machine learning model depends on one’s ability to understand all the details of the model including knowledge about different hyperparameters and how these parameters can be tuned in order to obtain the model with the best performance...
A robot puppet can learn to walk if it’s hooked up to human legs
Humans don’t need to have seen a set of stairs before in order to know what it is—or how to climb them. But for a robot, they can present an insurmountable problem. Getting robots to mimic how we manage to move around so effortlessly is one potential solution. That’s the premise of a study by researchers from the University of Illinois and MIT published in Science Robotics today...
Hugging Face: State-of-the-Art Natural Language Processing in ten lines of TensorFlow 2.0
Hugging Face is the leading NLP startup with more than a thousand companies using their library in production including Bing, Apple, Monzo. All examples used in this tutorial are available on Colab. The links are available in the corresponding sections...
Learning to See Moving Objects in the Dark.
Demo video and model...
3D BAT: A Semi-Automatic, Web-based 3D Annotation Toolbox for Full-Surround, Multi-Modal Data Streams
In this paper, we focus on obtaining 2D and 3D labels, as well as track IDs for objects on the road with the help of a novel 3D Bounding Box Annotation Toolbox (3D BAT)...
Coding habits for data scientists
In this article, we’ll share techniques for identifying bad habits that add to complexity in code as well as habits that can help us partition complexity...
Learning Data Manipulation for Augmentation and Weighting
Manipulating data, such as weighting data examples or augmenting with new instances, has been increasingly used to improve model training. Previous work has studied various rule- or learning-based approaches designed for specific types of data manipulation. In this work, we propose a new method that supports learning different manipulation schemes with the same gradient-based algorithm...
The First Step To Take When Looking For A Data Science Job
As someone new to the field, it all looks very chaotic and confusing to you. You want to find the right job for you and don't want to waste your time trying to apply to every single company advertising a job. Sadly, all the advice you are getting seems to be do a bunch of things that appear to be disconnected and sometimes even contradictory - network, blog, take these 4 MOOC's, talk to recruiters, don't talk to recruiters, participate in Kaggle competitions, do a bootcamp, join a data science fellowship program, etc... Given a limited amount of time and wanting to find the right opportunity, where do you start?...
Data Platform*
How to scale analytics across your entire company?
Dataform is a tool for managing data in your data warehouse, orchestrating complex SQL-based pipelines that take raw data and turn them into trusted datasets to power your company’s analytics. With Dataform your entire data team can collaborate on a single platform and manage advanced workflows. Version control, scheduling, data teats and development environments, all in SQL
Free hands-on consultation with our Data Architect Expert. They can help you import an existing project over and share best practices with you to help you get the most of the product. Book in an intro here
*Sponsored post. If you want to be featured here, or as our main sponsor, contact us!
Jobs
Data Scientist - Datadog - NYC
At Datadog, we’re on a mission to build the best monitoring platform in the world. We operate at high scale—trillions of data points per day—and high availability, providing always-on alerting, visualization, and tracing for our customers' infrastructure and applications around the globe.
Our engineering culture values pragmatism, honesty, and simplicity to solve hard problems the right way. We need you to design and build machine learning-powered products that help our customers learn from their data and make better decisions in real-time....
Want to post a job here? Email us for details >> team@datascienceweekly.org
Training & Resources
Keras Tuner documentation
Fully-featured, scalable, easy-to-use hyperparameter tuning for Keras & beyond...
μPlot: An exceptionally fast, tiny (~10 KB min) time series chart
μPlot is a fast, memory-efficient time series chart based on Canvas 2D; from a cold start it can create an interactive chart containing 150,000 data points in 40ms. In addition to fast initial render, the zooming and cursor performance is by far the best of any similar charting lib; at ~10 KB, it's likely the smallest and fastest time series plotter that doesn't make use of WebGL shaders or WASM, both of which have much higher startup cost and code size...
TensorFlowLite
A library for using TensorFlow Lite for Microcontrollers with Particle devices...
Books
The Lady Tasting Tea:
How Statistics Revolutionized Science in the Twentieth Century An insightful, revealing history of how mathematics transformed our world...
"I have taken courses in statistics, taught it many times and solved several statistical problems that have appeared in journals. But until I read this book, I never really thought about it in so deep and philosophical a manner..."
For a detailed list of books covering Data Science, Machine Learning, AI and associated programming languages check out our resources page.
P.S., Enjoy the newsletter? Please forward it to your friends and colleagues - we'd love to have them onboard :) All the best, Hannah & Sebastian