New data analysis competitions

How-to

Privacy

  • The Golden State Killer Is Tracked Through a Thicket of DNA, and Experts Shudder.

    Genetic testing services have become enormously popular with people looking for long-lost relatives or clues to hereditary diseases. Most never imagined that one day intimate pieces of their DNA could be mined to assist police detectives in criminal cases.

    Even as scientific experts applauded this week’s arrest of the Golden State Killer suspect, Joseph James DeAngelo, 72, some expressed unease on Friday at reports that detectives in California had used a public genealogy database to identify him. Privacy and ethical issues glossed over in the public’s rush to embrace DNA databases are now glaringly apparent, they said.

Tech

  • New AI Imaging Technique Reconstructs Photos with Realistic Results.

    Researchers from NVIDIA, led by Guilin Liu, introduced a state-of-the-art deep learning method that can edit images or reconstruct a corrupted image, one that has holes or is missing pixels.

    The method can also be used to edit images by removing content and filling in the resulting holes.

  • YouTube Says Computers Are Catching Problem Videos. As always, it will only take a while for someone to game the system, something will happen, like a video containing X, with X being something absolutely unacceptable, will rise to the top of the recommendations, and everybody will freak out, nobody is thinking of the children?, cats and dogs living together, and someone will tweak a parameter of the algorithm. Rinse, repeat.

    The vast majority of videos removed from YouTube toward the end of last year for violating the site’s content guidelines had first been detected by machines instead of humans, the Google-owned company said on Monday.

    YouTube said it took down 8.28 million videos during the fourth quarter of 2017, and about 80 percent of those videos had initially been flagged by artificially intelligent computer systems.

  • Nukes in the Age of AI.

    A new report from the RAND Corporation explains how artificial intelligence might affect the risk of nuclear war.

    Here's the report.

  • Using AI to trace leaking pipes.

    Almost 37% of the water pumped into Brazil’s supply network disappears, because of leaks, faulty meters returning an inaccurate reading or theft. That amounts to 6.5bn cubic metres in wasted water a year, worth some 8bn Brazilian real ($2.3bn) in forgone revenue, according to some estimates. Stattus4, based in Sorocaba, São Paulo state, reckons its portable leak-detection tool can not only indicate if water is going missing but also identify the different sorts of losses and where they might be.

    The company’s system, called Fluid, consists of a long hand-held probe on the end of which is a sensor. An operator places the sensor against an accessible area of pipework in, say, a culvert or under a manhole cover, for about 15 seconds to record vibrations in the water flowing through the local network. This sample is relayed to a smartphone app, which analyses the recording using a database stored on a cloud computer. This is a bit like how a smartphone music-recognition app, such as Shazam, works, says Marília Lara, a co-founder of the company. (The snatch of water music that Fluid listens to, however, also includes a range of frequencies which are inaudible to the human ear.)

    With enough samples, the company says it is possible to match sounds captured in one place to those of a leak a bit farther along the pipework, say. An illegal connection might be detected from the vibrations caused by an outflow where one is not supposed to be.

  • Generating fun Stack Exchange questions using Markov chains.

Visualizations


Data Links is a periodic blog post published on Sundays (specific time may vary) which contains interesting links about data science, machine learning and related topics. You can subscribe to it using the general blog RSS feed or this one, which only contains these articles, if you are not interested in other things I might publish.

Have you read an article you liked and would you like to suggest it for the next issue? Just contact me!