Month: May 2018

Our privacy policy & GDPR

Since the EU’s General Data Protection Regulation (GDPR) goes into effect today we lead with our own privacy policy, and our progress in incorporating GDPR requirements. We encourage you to read it. TL;DR: We do not yet claim complete compliance. But because we don’t depend on advertising, and don’t buy, sell, or share personal data to help others advertise, complying with GDPR is less daunting than it is for many others. Our privacy policy will be updated as we continue to add support for GDPR. Read More

How will the GDPR impact machine learning?

Since GDPR is bigger than Queen Bey, we’re all being bombarded by emails and articles about GDPR, ranging from the click-baity to the ill-informed to the wishful-thinking to the doomsday-wailing to the we-can-help, to the technically-bureaucratic, to the thoughtfully-analytical, und so weiter. We expect most of you by now have waded through enough to find a favorite or two to get a handle on the basics, so we’ve chosen two posts that deserve a careful read that you might have missed. First up, Andrew Burt, digs into the details and complexity of what GDPR means for machine learning. There is no simple answer, but Burt’s article is a good place to learn why. Read More

GDPR will pop the adtech bubble

It could happen. Note that Doc Searls is not talking about advertising in general, just adtech. But adtech is, shall we, say well-funded. Read More

And now a break from GDPR!​…

Martech simultaneous consolidation & expansion

How is this possible?…

… a more accurate view of martech consolidation cycles is that they are indeed happening  but they are happening on top of a wave of underlying software expansion… That underlying wave of software expansion seems to be overtaking

the natural business consolidation dynamics that are absolutely still happening at the same time. Read More

Also…

• More useful than Duplex… Maps walking navigation is Google’s most compelling use for AR yet via Techcrunch
• It’s not only about unstructured data… An Introduction to Deep Learning for Tabular Data via fast.ai
• Here’s who owns everything in Big Media today. Probably not for long. via Recode
• ‘A fun adventure, not a business’… The Weather Channel stopped publishing video on Facebook via Digiday

The Gilbane Advisor curates content for content, computing, and digital experience professionals. We focus on strategic technologies. We publish more or less twice a month except for August and December. See all issues

Annoying online ads do cost business

Nielsen Norman Group reports on a new research study involving 35 million Pandora users over 21 months. The study showed increased advertising caused a 2.8% reduction in use. As they point out, this is not a huge amount, and your mileage may vary.

What is significant is the convincing quantification. Nobody wants to have to defend a drop in customer activity. Read More

Predicting content attention and behavior

Content strategist Michael Andrews argues that “The biggest weakness in content strategy today is that it lacks predictive explanatory power. … To provide predictive explanatory power, content strategy guidelines should be based on empirical data that can be reproduced by others.”. Andrews summarizes, and points to, a new study presented at the 2018 World Wide Web Conference by Nir Grinberg of Northeastern University that provides some data and interesting analysis. The summary and Grinberg’s paper are both worthy of your time, and a must read if you’re a content strategist. Read More

How Axel Springer is getting consent for GDPR

They’ve been running some tests and are kindly sharing the results.

So far, the publisher’s readers are far more likely to give consent when they receive a fact-based static message, rather than a video message or one written in a tone that requests the readers’ support. Read More

Speech recognition systems vulnerable to adversarial attacks

Nicholas Carlini and David Wagner invented a novel attack against speech recognition AI. With the addition of an imperceptible amount of noise, the attack can trick speech-recognition systems into producing any output the attacker wants.

The Gradient’s Hugh Zhang points out that this kind of

attack is also a problem for other deep learning algorithms, for example in image recognition. Read More

Also…

• Philosophers Rule!… Ludwig Wittgenstein invented the emoji in 1938 via Quartz
• A relevant contemporary syllabus with links to excellent resources for corporations or others… Ethics in data science via University of Utah
• B2B marketers and content management: what’s working and what’s not via Digital Clarity Group
• 8 tentacles… Behind the curtain: How Google got media companies addicted via Axios

The Gilbane Advisor curates content for content, computing, and digital experience professionals. We focus on strategic technologies. We publish more or less twice a month except for August and December. See all issues