Every two weeks, we find the most interesting data science links from around the web and collect them in Data Science Briefings, the DataMiningApps newsletter. Subscribe now for free if you want to be the first to get up to speed on interesting resources.
- WaveNet: A Generative Model for Raw Audio
All the rave this week: another breakthrough from Deepmind: “This post presents WaveNet, a deep generative model of raw audio waveforms. We show that WaveNets are able to generate speech which mimics any human voice and which sounds more natural than the best existing Text-to-Speech systems, reducing the gap with human performance by over 50%.”
- AI Can Recognize Your Face Even If You’re Pixelated
Researchers at the University of Texas at Austin and Cornell Tech say that they’ve trained a piece of software that can undermine the privacy benefits of standard content-masking techniques like blurring and pixelation by learning to read or see what’s meant to be hidden in images.
- Getting a sense of statistics… by eating them
How do you convey complicated data in a way that’s easier to understand? A group of researchers think the answer may lie in our tastebuds.
- The reward engineering problem in reinforcement learning
Today we usually train reinforcement learning agents to perform narrow tasks with simple goals. We may eventually want to train RL agents to behave “well” in open-ended environments where there is no simple goal. In order to evaluate these agents, the question is: how do we carry out the evaluation, so that the optimal strategy for A is to also make “good” decisions?
- An AI system at Houston Methodist Hospital read breast X-rays 30x faster than doctors, with 20% greater accuracy
Artificial intelligence researchers at Houston Methodist Hospital have developed computer software that could make the dreaded mammogram callback quicker and more reliable.
- Need Some AI? Yeah, There’s a Marketplace for That
Diego Oppenheimer is worried that the Googles and the Facebooks will dominate the world of artificial intelligence. Oppenheimer and his startup, Algorithmia, are doing their part in the battle against AI hegemony. Algorithmia is what Oppenheimer calls an open marketplace for algorithms—code that companies and developers can use to beef up their websites and apps.
- When your boss is an algorithm
In the gig economy, companies such as Uber and Deliveroo manage workers via their phones. But is this liberating or exploitative?
- War-Algorithm Accountability (paper)
“In this briefing report, we introduce a new concept — war algorithms — that elevates algorithmically-derived “choices” and “decisions” to a, and perhaps the, central concern regarding technical autonomy in war. We thereby aim to shed light on and recast the discussion regarding “autonomous weapon systems.””
- Self-Driving Cars Can Learn a Lot by Playing Grand Theft Auto
Hyper-realistic computer games may offer an efficient way to teach AI algorithms about the real world.
- R for Data Science
The draft of the new book by R legends Garrett Grolemund and Hadley Wickham is finished and can be read online. Great resource for how modern R in data science should be done!
- A Study of Generative Algorithms
Not really pure data science, but a series of beautiful algorithmic showcases nonetheless.
- How to train your #NeuralNetwork for Wine tasting?
When it comes to classification of wine, the practice is quite varied based on region of origin and time. It is one of the most tasteful traditions which is also protected by law of its own in certain regions. Is it possible to teach something about the classification of different variety of wines to Neural Networks?
- The Neural Network Zoo
With new neural network architectures popping up every now and then, it’s hard to keep track of them all. Knowing all the abbreviations being thrown around (DCIGN, BiLSTM, DCGAN, anyone?) can be a bit overwhelming at first. This site presents a cheat sheet containing many of those architectures. Fantastic resource!
- Beware of the gaps in Big Data
As we entrust ever more of our lives to ‘big data’, how can we protect against the gaps and mistaken assumptions used to handle the information?
- Telco churn prediction with R and H2O
This showcase presents how easy it is to use H2O library to build high quality predictive models.
- 360 video stabilization: A new algorithm for smoother 360 video viewing by Facebook
As more people are shooting more immersive videos in real-life scenarios and sharing these moments with their friends, Facebook has developed a new stabilization technology custom-designed for 360 video.
- The Extraordinary Link Between Deep Neural Networks and the Nature of the Universe
Nobody understands why deep neural networks are so good at solving complex problems. Now physicists say the secret is buried in the laws of physics.
Make visualisations using only markdown using a simple declarative markup like you would write code. Interesting idea.
- Jupyter (IPython) notebooks features
Familiar already to most data scientists, but still a good overview.