Google Gives Third Day Keynote at Pubcon
November 1, 2016
Technology conferences are the thing to do when you want to launch a product, advertise a new business, network, or get a general consensus about the tech industry. There are multiple conferences revolving around different aspects in the tech industry held each month. In October 2016, Pubcon took place in Las Vegas, Nevada and they had a very good turn out. The thing that makes a convention, though, is the guests. Pubcon did not disappoint as on the third day, Google’s search expert Gary Illyes delivered the morning keynote. (Apparently, Illyes also hold the title Chief of Sunshine and Happiness at Google). Outbrain summed up the highlights of Pubcon 2016’s third day in “Pubcon 2016 Las Vegas: Day 3.”
Illyes spoke about search infrastructure, suggesting that people switch to HTTPS. His biggest push for HTTPS was that it protected users from “annoying scenarios” and it is good for UX. Google is also pushing for more mobile friendly Web sites. It will remove “mobile friendly” from search results and AMP can be used to make a user-friendly site. There is even bigger news about page ranking in the Google algorithm:
Our systems weren’t designed to get two versions of the same content, so Google determines your ranking by the Desktop version only. Google is now switching to a mobile version first index. Gary explained that there are still a lot of issues with this change as they are losing a lot of signals (good ones) from desktop pages that are don’t exist on mobile. Google created a separate mobile index, which will be its primary index. Desktop will be a secondary index that is less up to date.
As for ranking and spam, Illyes explained that Google is using human evaluators to understand modified search better, Rankbrain was not mentioned much, he wants to release the Panda algorithm, and Penguin will demote bad links in search results. Google will also release “Google O for voice search.
It looks like Google is trying to clean up search results and adapt to the growing mobile market, old news and new at the same time.
Whitney Grace, November 1, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
How to Find an Email Address
October 27, 2016
Like any marketers, search engine optimizers must reach out to potential clients, and valid email addresses are important resources. Now, Search Engine Journal explains “How to Find Anyone’s Email Address in 60 Seconds or Less.” Anyone’s, really? Perhaps that’s a bit of an exaggeration.
SEO pro, Joshua Daniels discusses six techniques to quickly find an email address. He writes:
If you’re a specialist in SEO or link acquisition, then you’ll know that generic email addresses are as much use as a chocolate fireguard when it comes to outreach. You need to develop personal connections with influencers, regardless of whether you work in PR or SEO, it’s always the same. But finding the right person’s email address can be a draining, time-consuming task. Who has time for that?
Well, actually, it’s not so difficult, or time-consuming. In this post, I’m going to walk you through the exact step-by-step process our agency uses to find (almost) anyone’s email address, in 60 seconds or less!
For each suggestion, Daniels provides instructions, most with screen shots. First, he recommends LinkedIn’s search function paired with Email Hunter, a tool which integrates with the career site. If that doesn’t work, he says, try a combination of the Twitter analyzer Followerwonk and corporate-email-finder Voila Norbert.
The article also suggests leveraging Google’s search operators with one of these formats: [site:companywebsite.com + “name” + contact] or [site:companywebsite.com + “name” + email]. To test whether an email address is correct, verify it with MailTester, and to target someone who posts on Twitter, search the results of All My Tweets for keywords like “email” or “@companyname.com”. If all else fails, Daniels advises, go old school—“… pick up the phone and just ask.”
Cynthia Murrell, October 27, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Partnership Aims to Establish AI Conventions
October 24, 2016
Artificial intelligence research has been booming, and it is easy to see why— recent advances in the field have opened some exciting possibilities, both for business and society as a whole. Still, it is important to proceed carefully, given the potential dangers of relying too much on the judgement of algorithms. The Philadelphia Inquirer reports on a joint effort to develop some AI principles and best practices in its article, “Why This AI Partnership Could Bring Profits to These Tech Titans.” Writer Chiradeep BasuMallick explains:
Given this backdrop, the grandly named Partnership on AI to Benefit People and Society is a bold move by Alphabet, Facebook, IBM and Microsoft. These globally revered companies are literally creating a technology Justice League on a quest to shape public/government opinion on AI and to push for friendly policies regarding its application and general audience acceptability. And it should reward investors along the way.
The job at hand is very simple: Create a wave of goodwill for AI, talk about the best practices and then indirectly push for change. Remember, global laws are still obscure when it comes to AI and its impact.
Curiously enough, this elite team is missing two major heavyweights. Apple and Tesla Motors are notably absent. Apple Chief Executive Tim Cook, always secretive about AI work, though we know about the estimated $200 million Turi project, is probably waiting for a more opportune moment. And Elon Musk, co-founder, chief executive and product architect of Tesla Motors, has his own platform to promote technology, called OpenAI.
Along with representatives of each participating company, the partnership also includes some independent experts in the AI field. To say that technology is advancing faster than the law can keep up with is a vast understatement. This ongoing imbalance underscores the urgency of this group’s mission to develop best practices for companies and recommendations for legislators. Their work could do a lot to shape the future of AI and, by extension, society itself. Stay tuned.
Cynthia Murrell, October 24, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Picking Away at Predictive Programs
October 21, 2016
I read “Predicting Terrorism From Big Data Challenges U.S. Intelligence.” I assume that Bloomberg knows that Thomson Reuters licenses the Palantir Technologies Metropolitan suite to provide certain information to Thomson Reuters’ customers. Nevertheless, I was surprised at some of the information presented in this “real” journalism write up.
The main point is that numerical recipes cannot predict what, when, where, why, and how bad actors will do bad things. Excluding financial fraud, which seems to be a fertile field for wrong doing, the article chases the terrorist angle.
I learned:
- Connect the dots is a popular phrase, but connecting the dots to create a meaningful picture of bad actors’ future actions is tough
- Big data is a “fundamental fuel”
- Intel, PredPol, and Global Intellectual Property Enforcement Center are working in the field of “predictive policing”
- The buzzword “total information awareness” is once again okay to use in public
I highlighted this passage attributed too a big thinker at the Brennan Center for Justice at NYU School of Law:
Computer algorithms also fail to understand the context of data, such as whether someone commenting on social media is joking or serious,
Several observations:
- Not a single peep about Google Deep Mind and Recorded Future, outfits which I consider the leaders in the predictive ball game
- Not a hint that Bloomberg was itself late to the party because Thomson Reuters, not exactly an innovation speed demon, saw value in Palantir’s methods
- Not much about what “predictive technology” does.
In short, the write up delivers a modest payload in my opinion. I predict that more work will be needed to explain the interaction of math, data, and law enforcement. I don’t think a five minute segment with talking heads on Bloomberg TV won’t do it.
Stephen E Arnold, October 21, 2016
Pattern of Life Analysis to Help Decrypt Dark Web Actors
October 18, 2016
Google funded Recorded Future plans to use technologies like natural language processing, social network analysis and temporal pattern analysis to track Dark Web actors. This, in turn, will help security professionals to detect patterns and thwart security breaches well in advance.
An article Decrypting The Dark Web: Patterns Inside Hacker Forum Activity that appeared on DarkReading points out:
Most companies conducting threat intelligence employ experts who navigate the Dark Web and untangle threats. However, it’s possible to perform data analysis without requiring workers to analyze individual messages and posts.
Recorded Future which deploys around 500-700 servers across the globe monitors Dark Web forums to identify and categorize participants based on their language and geography. Using advanced algorithms, it then identifies individuals and their aliases who are involved in various fraudulent activities online. This is a type of automation where AI is deployed rather than relying on human intelligence.
The major flaw in this method is that bad actors do not necessarily use same or even similar aliases or handles across different Dark Web forums. Christopher Ahlberg, CEO of Recorded Future who is leading the project says:
A process called mathematical clustering can address this issue. By observing handle activity over time, researchers can determine if two handles belong to the same person without running into many complications.
Again, researchers and not AI or intelligent algorithms will have to play a crucial role in identifying the bad actors. What’s interesting is to note that Google, which pretty much dominates the information on Open Web is trying to make inroads into Dark Web through many of its fronts. The question is – will it succeed?
Vishal Ingole, October 18, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
A Literary Magazine by a Machine?
October 14, 2016
Literary magazines are a great way to read short stories, the latest poetry, and other compelling pieces by a variety of authors. What if those authors are machines? CuratedAI is the first literary magazine written by machines for human readers. Computers are presented as sterile, uncreative items, but technology programmed with machine learning and content curation can actually write some decent pieces.
Here is the magazine’s mission statement:
“CuratedAI is a literary magazine with a twist– all stories and poems are generated by machines using the tricks of the Artificial Intelligence trade. Editing, for now, is still the domain of us humans, but we aim to keep our touch as light as possible.”
Poetry is a subjective literary form and perhaps the most expressive. It allows writers to turn words into art and stray away from standard language rules. In other words, it is the perfect form for computers. They insert adjectives wherever the algorithm states and the sentences do not always make sense.
Prose, on the other mouse, is not its best form. The stories read like a bad Internet translation from Japanese to English. It will be some time before computers are writing comprehensible novels, at least for some of them. Machine learning was used in Japan for a novel writing contest and the machine that wrote the book, actually won a prize. So machine cans write prize-winning literature.
However, no one can program imagination…not yet anyway.
Whitney Grace, October 14, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Malware with Community on the Dark Web
October 14, 2016
While Mac malware is perhaps less common than attacks designed for PC, it is not entirely absent. The Register covers this in a recent article, EasyDoc malware adds Tor backdoor to Macs for botnet control. The malware is disguised as a software application called EasyDoc Converter which is supposed to be a file converter but does not actually perform that function. Instead, it allows hackers to control the hacked mac via Tor. The details of the software are explained as follows,
The malware, dubbed Backdoor.MAC.Eleanor, sets up a hidden Tor service and PHP-capable web server on the infected computer, generating a .onion domain that the attacker can use to connect to the Mac and control it. Once installed, the malware grants full access to the file system and can run scripts given to it by its masters. Eleanor’s controllers also uses the open-source tool wacaw to take control of the infected computer’s camera. That would allow them to not only spy on the victim but also take photographs of them, opening up the possibility of blackmail.
A Computer World article on EasyDoc expands on an additional aspect of this enabled by the Dark Web. Namely, there is a Pastebin agent which takes the infected system’s .onion URL, encrypts it with an RSA public key and posts it on Pastebin where attackers can find it and use it. This certainly seems to point to the strengthening of hacking culture and community, as counterintuitive of a form of community, it may be to those on the outside.
Megan Feil, October 14, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Autonomy Founder Lynch Snags Bloomberg Profile
October 7, 2016
I read “Former Autonomy CEO Mike Lynch Is Running a Hands-on Fund While He Battles HP in Court.” I have a modest file of open source information about the Hewlett Packard purchase of Autonomy, the company which sparked a shift in search and information access. Years ago, I even met some Autonomy professionals and performed a small task for which I was compensated in a modest way. Autonomy and its Integrated Data Operating Layer / Digital Reasoning Engine remain important milestones. Autonomy was one of the first firms to apply fancy math to thorny problems in making sense of unstructured information. Today, Dr. Lynch’s approach informs many of the smart software companies capturing headlines about online translation to predicting which team will win a football game.
Bloomberg Businessweek’s write up focuses on Dr. Michael Lynch’s life after his split from HP. That Sillycon Valley icon bought Autonomy and seems to have sold the property on again. American real estate television programs call this process a “flip.” The idea is to buy low and sell high. HP’s approach to a “flip” is a bit different. HP bought high and seems to have sold low. That’s why I use the term “Sillycon Valley” to describe the executive methods associated with San Francisco and environs.
The write up tells me:
His $1 billion Europe-focused firm, Invoke Capital, looks like a cross between the Carlyle Group and Y Combinator.
His business approach is described this way:
“We want an unfair advantage. The internal saying is, always take a gun to a knife fight,” Lynch says. As for Invoke’s pitch to startups: “There’s a little bit of an element of the Spice Girls. We can bring people together.”
The article focuses on Luminance, a company with smart software for legal eagles. The company I find most interesting is Dr. Lynch’s Darktrace.
HP Enterprise is suing Dr. Lynch. The legal shoot out will take place in 2017. The write up correctly points out that Dr. Lynch may show up in court to explain that he generated more revenue, more value, and more saleable software than HP did with Autonomy’s technology.
The Bloomberg report does not make these points:
- Dr. Lynch’s products tap his knowledge of numerical recipes. When applied in an informed manner, these methods are useful and have applicability to a number of business problems. HP bought Autonomy and seems to have lacked the expertise or motivation or business acumen to output big money
- HP decided to buy Autonomy and then suffered buyer’s remorse in the midst of management turmoil and efforts to prevent Hewlett Packard from becoming an also-ran in the Sillycon Valley growth race
- HP wants to prove Dr. Lynch fooled them even though HP’s professionals and consultants performed due diligence prior to buying Autonomy. Perhaps HP should look at its systems and methods?
Worth monitoring even if an observer does not understand Bayesian, Laplacian, Monte Carlo, and Markovian methods. My hunch is that HP’s lawyers may find these mathematical methods helpful in determining the odds for certain events. When one feeds data about Dr. Lynch’s current hot start ups into the model, the predictive output suggests more revenue and innovation for Dr. Lynch. HP, on the other hand, may receive a different set of probabilities.
If you are not familiar with the Autonomy information access system, I have made available, a free report about Autonomy. You may access it at my Xenky.com archive.
Stephen E Arnold, October 7, 2017
Quote to Note: Microsoft CEO on Tay
October 3, 2016
I circled this quote to note in “Microsoft CEO Satya Nadella on Artificial Intelligence, Algorithmic Accountability, and What He Learned from Tay.” Here’s the statement about Tay, the chat bot which developed some interesting characteristics:
One of my biggest learnings from [chatbot] Tay was that you need to build even AI that is resilient to attacks. It was fascinating to see what happened on Twitter, but for instance we didn’t face the same thing in China. Just the social conversation in China is different, and if you put it in the US corpus it’s different. Then, of course, there was a concerted attack.
Yep, algorithmic bias. Perhaps in the rush to make a marketing splash some steps were hop scotched?
Who was responsible for the mistake?
We were [building Tay] as a prototype to learn. Right now, given the media cycle that we have, there is no distinction between a prototype and a released product.
The media. I knew it.
Stephen E Arnold, October 3, 2016
MCMC: No, Not a Musical Trend
September 28, 2016
If you wonder why a Grand Canyon exists between the people who offer point-and-click statistical analysis and the folks who “do” the math, you will want to read “Markov Chain Monte Carlo Without All…” When you read the source document, keep in mind that selecting an icon and generating a report is like using an automatic teller machine. Punching buttons delivers an output. The inner workings of the system are not visible. User friendly, embedded business intelligence systems are like the chrome trim on a door stop.
Stephen E Arnold, September 28, 2016
-
- Subscribe to Beyond Search
Feature archive
News archive
-
