Monday, May 20, 2019


Maybe Facebook could buy Finland?
Finland is winning the war on fake news. What it’s learned may be crucial to Western democracy
Finland has faced down Kremlin-backed propaganda campaigns ever since it declared independence from Russia 101 years ago. But in 2014, after Moscow annexed Crimea and backed rebels in eastern Ukraine, it became obvious that the battlefield had shifted: information warfare was moving online.
Toivanen, the chief communications specialist for the prime minister’s office, said it is difficult to pinpoint the exact number of misinformation operations to have targeted the country in recent years, but most play on issues like immigration, the European Union, or whether Finland should become a full member of NATO (Russia is not a fan).




AI is too easily fooled.
How we might protect ourselves from malicious AI
We’ve touched previously on the concept of adversarial examples—the class of tiny changes that, when fed into a deep-learning model, cause it to misbehave. In March, we covered UC Berkeley professor Dawn Song’s talk at our annual EmTech Digital conference about how she used stickers to trick a self-driving car into thinking a stop sign was a 45-mile-per-hour sign, and how she used tailored messages to make a text-based model spit out sensitive information like credit card numbers. In April, we similarly talked about how white hat hackers used stickers to confuse Tesla Autopilot into steering a car into oncoming traffic.
… A new paper from MIT now points toward a possible path to overcoming this challenge. It could allow us to create far more robust deep-learning models that would be much harder to manipulate in malicious ways. To understand its significance, let’s first review the basics of adversarial examples.
It showed this by identifying a rather interesting property of adversarial examples that helps us grasp why they’re so effective. The seemingly random noise or stickers that trigger misclassifications are actually exploiting very precise, minuscule patterns that the image system has learned to strongly associate with specific objects. In other words, the machine isn’t misbehaving when it sees a gibbon where we see a panda. It is indeed seeing a pattern of pixels, imperceptible to humans, that occurred far more often in the gibbon photos than panda photos during training.




Another huge list. Browsers and their targeting, specialized tools.
New on LLRX – Online Research Browsers 2019
Via LLRX Online Research Browsers 2019 Marcus Zillman’s guide highlights multifaceted browser alternatives to mainstream search tools that researchers may regularly use by default. There are many reliable yet underutilized applications that facilitate access to and discovery of subject matter specific documents and sources. Free applications included here also offer collaboration tools, resources to build and manage repositories, to employ data visualization, to create and apply metadata management, citations, bibliographies, document discovery and data relationship analysis.



No comments: