Google DeepMind AI Project Makes Progress
July 25, 2016
For anyone following the development of artificial intelligence, I recommend checking out the article, “How Google Plans to Solve Artificial Intelligence” at MIT Technology Review. The article delves into Google’s DeepMind project, an object of renewed curiosity after its AlphaGo software bested the human world champion of the ancient game Go in March.
This Go victory is significant, because it marks progress beyond the strategy of calculating different moves’ possible outcomes; the game is too complex for that established approach (though such calculations did allow IBM’s DeepBlue to triumph over the world chess champion in 1997). The ability to master Go has some speaking of “intuition” over calculation. Just how do you give software an approximation of human intuition? Writer Tom Simonite tells us:
“Hassabis believes the reinforcement learning approach is the key to getting machine-learning software to do much more complex things than the tricks it performs for us today, such as transcribing our words, or understanding the content of photos. ‘We don’t think just observing is enough for intelligence, you also have to act,’ he says. ‘Ultimately that’s the only way you can really understand the world.’”
“DeepMind’s 3-D environment Labyrinth, built on an open-source clone of the first-person-shooter Quake, is designed to provide the next steps in proving that idea. The company has already used it to challenge agents with a game in which they must explore randomly generated mazes for 60 seconds, winning points for collecting apples or finding an exit…. Future challenges might require more complex planning—for example, learning that keys can be used to open doors. The company will also test software in other ways, and is considering taking on the video game Starcraft and even poker. But posing harder and harder challenges inside Labyrinth will be a major thread of research for some time, says Hassabis. “It should be good for the next couple of years,” he says.”
The article has a video of DeepMind’s virtual labyrinth you can check out, if you’re curious. (It looks very much like an old Windows screen saver some readers may recall.) Simonite tells us that AI firms across the industry are watching this project carefully. He also points to some ways DeepMind is already helping with real-world problems, like developing training software with the U.K.’s National Health Service to help medical personnel recognize commonly missed signs of kidney problems.
See the article for much more about Google’s hopes and plans for DeepMind. Simonite concludes by acknowledging the larger philosophical and ethical concerns around artificial intelligence. We’re told DeepMind has its own “internal ethics board of philosophers, lawyers, and businesspeople.” I think it is no exaggeration to say these folks, whom Google indicates it will name someday soon, could have great influence over the nature of our future technology. Let us hope Google chooses wisely.
Cynthia Murrell, July 25, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
There is a Louisville, Kentucky Hidden Web/Dark Web meet up on July 26, 2016. Information is at this link: http://bit.ly/29tVKpx.
You Do Not Tay?
July 25, 2016
The article titled Microsoft CaptionBot: AI Image Guessing App Really Isn’t Sure Who Barack Obama Is on International Business Times assesses Microsoft’s latest attempt at AI following the catastrophic Twitter robot Tay which quickly “learned” and repeated some pretty darn offensive ideas about Hitler and Obama. The newly released version named CaptionBot is more focused on image descriptions. The article states,
“Users are asked to upload any photo to the site, then Microsoft’s AI system attempts to describe what is in the image. The system can recognise celebrities and understands the basics of image composition but…, it isn’t yet perfect… You know when you recognise someone, but can’t quite put your finger on who it is? Caption Bot doesn’t do that, it just fails to even describe what a photo of Barack Obama is, never mind who he might be.”
From the examples, it is clear that while CaptionBot is much better at understanding and defining objects than people, objects often create difficulty as well. An image of a yellow vehicle from Cars was described (without confidence) as a white toilet next to a yellow building. To be sure, if you stare at the image long enough, the toilet shape emerges.
Chelsea Kerwin, July 25, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
There is a Louisville, Kentucky Hidden Web/DarkWeb meet up on July 26, 2016. Information is at this link: http://bit.ly/29tVKpx.2
Sentiment Analysis: Now Embrace Mood Mapping
July 24, 2016
I love the buzzword game. Move enterprise search to sentiment analysis. Then another player nudges forward “mood mapping.” I expect to hear “checkmate,” but I only notice the odor of baloney.
Navigate to “Move Over Sentiment Analysis, Mood Mapping Is Here.” I learned or was exposed to this passage:
“Sentiment analysis has long been the standard approach for understanding how consumers feel about a brand. While this approach sticks to aggregating positive, negative and neutral sentiments; in a world of colors, that’s like having only black, white and grey. With Mood Mapping, we have introduced a next-generation algorithm that helps marketers understand how consumers feel about a particular brand and how they are likely to act on their feelings, something that brand marketers care for very deeply,” Amarpreet Kalkat, co-founder-CEO, Frrole, told [the author, maybe?].
And the company is an award winner. I learned:
Founded in 2014, Frrole, a Microsoft Ventures Accelerator-incubated company, had raised an angel round of $2,45,000 in 2014. The profitable company expects to hit $1 million in revenue run-rate by the end of this year. The start-up was recognized as ‘Marketing Technologist of the Year’ at the Big Data and Analytics Summit 2016.
I think at the pre school near our home, every child gets an award whether a victor in a competition or not. Perhaps Microsoft will integrate mood mapping with LinkedIn in Office 365. That will be exciting when Word’s numbering function fails to count.
Stephen E Arnold, July 24, 2016
Bing Becomes More Like Telex
July 23, 2016
Decades ago, probably around 1980, I met with a person in New York City. The purpose of the meeting was to talk about value added information provided to users of the Telex system. (If you want a round up of Telex, let Wikipedia do it at this link.)
At that meeting, I saw the information available on the Telex “menu” for operators. There was a list of pop songs as I recall. Other “intellectual” goodies were the day’s astrological predictions, movies, and some star info. The demonstration consumed Telex minutes and, therefore, produced money for the outfit with which I was meeting. I knew that value added information like the content for which I responsible decades ago was a loser, non starter, dead dog, and of zero interest to this big company. The Telex provided information was breezy and designed to keep bored Telex operators amused when not typing out messages of grave import. It was, in a word, fluff. But it made money until the Telex world was vaporized by digital flows incompatible with double digit baud text delivery.
I thought of this demo when I read “Bing’s Home Page Gets Smart with Trivia, Quizzes & Polls.” The Bing innovation is a lowest common denominator function. Like the Telex information, it appeals to those with little to do. The intellectual payload is close to zero.
The write up, for understandable reasons, is not in the business of reminding the new Microsoft that it is imitating a Telex system’s content. I learned:
Senior managing editor Kristen Kennedy and senior program manager Vinay Krishna said they want Bing’s homepage to be a source of inspiration for millions and an entry point to learn more about the world.
Learning can be fun. What’s your sign? How is that horoscope’s accuracy today, gentle reader? What is the lowest common search denominator? Answer: Bing maybe?
Stephen E Arnold, July 23, 2016
Palantir Thiel: An If Then Chess Move?
July 22, 2016
I read “The Peter Principle: Why Thiel’s GOP Convention Speech Will Be about Him and Not about Silicon Valley.” Interesting write up but I think the “about” part is possibly incorrect. I think the speech just might have been about Palantir, procurement, and displacing the traditional US government defense contractors. The stakes are not ego; the stakes are hundreds of millions in technology business. Silicon Valley and money. It is possible that some Palantir-think informs the political enthusiasms of Mr. Thiel.
In the write up, I noted this passage.
The eccentric investor is not like anyone else in tech.
That is okay. Just like his chess moves, the expected approach may not disrupt an strong opponent’s game. But for those who are not chess masters, the unusual or less probable maneuver can open some doors.
I also circled this paragraph:
He’s been vocal about that at many political gatherings here over the years, including one where those present said he even freaked Tea Party favorite Rep. Paul Ryan out with his thoughts on the need to tear down government as much as possible. It’s not just regulation that Thiel hates, it’s the whole stinking mess, and he has not been shy about saying so. Thus, most people who know him usually give you an oh-that-guy look, shrug their shoulders and move on to easier topics. And many of those who like him personally and call him a “dear friend” are even perplexed, going as far as not talking to him of late because of his Trump support.
The statement is almost edging toward a motivation which I think may be operating. How is this for a possibility:
Palantir wants to provide systems and software to the US government. Mr. Thiel becomes an insider to Mr. Trump, who, surprising many, becomes the president of the good, old USA. As an insider, Mr. Thiel can bring some pressure to bear on those in the government who do not demonstrate the type of enthusiasm for Palantir Technologies which are warranted. The irrational decision making and the failure to follow procurement rules might become a focal point for a Trump-centric administration. Checkmate, US Army. Palantir wins.
There are many moves in chess. This might be one of them with Mr. Thiel looking four or five moves ahead. Critics of Mr. Thiel just might be overlooking an important driver for his Trump support — Palantir’s financial and procurement future. If Mr. Trump wins in November 2016, then Palantir can try to bring rationality to Palantir procurements by certain Federal entities. Just a thought, gentle reader. Just a mental chimera like opening up lines of attack for a bishop and a queen in chess.
Stephen E Arnold, July 22, 2016
DuckDuckGo: Filtering
July 22, 2016
I read “Is DuckDuckGo.com Partially Enforcing the “Celebrity Threesome Injunction“? The point of the write up is that information is filtered from search systems, including the privacy-centric system DuckDuckGo.com. I assume the queries summarized in the write up are spot on. If accurate, one cannot search that which is not in an index. That’s helpful for those who want to be thorough. It is also helpful for those who find themselves the subject of write ups already published and want to keep the links out of a search system’s results page. With folks loving the mobile research experience, who would know? The more interesting question, “Does anyone care?” A good example is the “artist” whose work disappeared from the Alphabet Google thing’s Blogger system. See “Google Deletes Artist’s Blog and a Decade of His Work along with It.” Back ups are good if not filtered by a helpful cloud service. Where did my music go anyway?
Stephen E Arnold, July 22, 2016
Alphabet Google Is Busy Reinventing
July 22, 2016
From Forbes in India (“Sundar Pichai to Reinvent Google with a Heavy Dose of Artificial Intelligence” which may require a proxy maneuver due to the digitally with it Forbes) or Switzerland (“Google’s New Research Lab in Zurich Is Inventing the Future of Search”) — the Alphabet Google thing is trying to reinvent search.
There you go: Stark evidence that Google information retrieval system is deeply flawed. The electric car does not reinvent the car. But search has to reinvent search.
This is a big and probably futile job. My view is that search is an evolutionary beastie. Incremental innovations from research labs, one man band coders, and start ups with one good idea and couple of crazed investors do the job.
Google itself was a roll up of ideas from IBM Almaden (hell, Jon Kleinberg), AltaVista (hello, Jeff Dean, Simon Tong, and Sanjay Ghemawat), and the fumble bumbles of folks at precursors (hello, AskJeeves and Lycos).
The India angle states:
Think of it as Search 3.0—a new, interactive way to communicate with Google itself. With it you’ll be able to order a ticket, book a flight, play music, schedule a task, reply to a message; the Google assistant might even write it for you. It might prompt you to order flowers ahead of Mother’s Day or to pack for your upcoming trip, and it might be able to pick up an earlier conversation from where you left off. In other words, it will be there, ready to help, in your phone, your speakers, your television, your car, your watch and eventually everywhere. “You are trying to go about your day, and in an ambient way, things are there to help you,” Pichai says. Making sure this assistant lives up to its full potential will take years, and building it will be harder than it was for Page and co-founder Sergey Brin to create search itself. Adds Pichai: “In every dimension, it is more ambitious.”
Yep, ambitious.
From the Swiss side:
he new team has a distinct goal: to invent the future of Search, a voice-activated, human-like entity that can answer any query intelligently. “We are building the ultimate assistant. In two years, you can expect Google to become a personal life assistant across multiple surfaces, including your phone, Google Home, even cars,” Mogenet [Google wizard] said. Some of Google’s best-known products are already shaped by machine learning, the ability of computers to spot patterns in large datasets and learn by example. For instance, Google Photos uses it to understand the content of an image. This means you could search for “cardigan corgi” or “passport” or “birthday celebrations 2014” and the app will bring up the relevant photos.
There you go. Reinvent.
The challenge is to find a way to avoid the stagnation which seems to befall certain types of high technology outfits. Do you use your DEC Rainbow today?
I love the Google. It is just super. The problem is that as it has concentrated traffic, it has left itself unable to respond to opportunities such as those identified by Facebook and Amazon. By the way, both of these outfits face some challenges as well.
The investment in search will benefit some folks. But how likely is it that Google will come up with an “innovation” that matters. I think that when octopus companies do something — whether it is good or bad — it is easy to define whatever happens as success.
The problem is that information returned from Google is often off point. When I run queries for documents I have in my hand, I cannot find them without jumping through hoops. I documented this with a Dark Web paper from Denmark in this blog. Homonyms give the Google fits. Even though my search history is available to Mother Google, the system is tone deaf for my queries. When I look for certain information, the data are often disappeared. I noticed that indexing of pastesites, PDF files, and PowerPoint presentations has become laughable.
Innovation is more than a public relations campaign. How do I know? Google’s marketing is starting to remind me of IBM Watson. You know Watson, the revolutionary information access system from Big Blue. Yep, innovation.
Stephen E Arnold, July 22, 2016
Oracle v Google Copyright Trial in Progress
July 22, 2016
The battle between Google and Oracle over Android’s use of Java has gone to federal court, and the trial is expected to conclude in June. CBS San Francisco Bay Area reports, “Former Google CEO Testifies in Oracle-Google Copyright Trial.” The brief write-up reveals the very simple defense of Eric Schmidt, who was Google’s CEO while Android was being developed (and is now CEO of Google’s young parent company, Alphabet): “We believed our approach was appropriate and permitted,” he stated.
Java was developed back in the ‘90s by Sun Microsystems, which was bought by Oracle in 2010. Google freely admits using Java in the development of Android, but they assert it counts as fair use—the legal doctrine that allows limited use of copyrighted material if it is sufficiently transformed or repurposed. Oracle disagrees, though Schmidt maintains Sun Microsystems saw it his way back in the day. The article tells us:
“Schmidt told the jury that when Google was developing Android nine years ago, he didn’t believe the company needed a license from Sun for the APIs. “We believed our approach was appropriate and permitted,” he said.
“Under questioning from Google attorney Robert Van Nest, Schmidt said that in 2007, Sun’s chief executive officer Jonathan Schwartz knew Google was building Android with Java, never expressed disapproval and never said Google needed a license from Sun.
“In cross-examination by Oracle attorney Peter Bicks, Schmidt acknowledged that he had said in 2007 that Google was under pressure to compete with the Apple Inc.’s newly released iPhone.”
Yes it was, the kind of pressure that can erode objectivity. Did Google go beyond fair use in this case? The federal court will soon decide.
Cynthia Murrell, July 22, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
There is a Louisville, Kentucky Hidden Web/Dark
Web meet up on July 26, 2016.
Information is at this link: http://bit.ly/29tVKpx.
Meet the Company Selling Our Medical Data
July 22, 2016
A company with a long history is getting fresh scrutiny. An article at Fortune reports, “This Little-Known Firm Is Getting Rich Off Your Medical Data.” Writer Adam Tanner informs us:
“A global company based in Danbury, Connecticut, IMS buys bulk data from pharmacy chains such as CVS , doctor’s electronic record systems such as Allscripts, claims from insurers such as Blue Cross Blue Shield and from others who handle your health information. The data is anonymized—stripped from the identifiers that identify individuals. In turn, IMS sells insights from its more than half a billion patient dossiers mainly to drug companies.
“So-called health care data mining is a growing market—and one largely dominated by IMS. Last week, the company reported 2015 net income of $417 million on revenue of $2.9 billion, compared with a loss of $189 million in 2014 (an acquisition also boosted revenue over the year). ‘The outlook for this business remains strong,’ CEO Ari Bousbib said in announcing the earnings.”
IMS Health dates back to the 1950s, when a medical ad man sought to make a buck on drug-sales marketing reports. In the 1980s and ‘90s, the company thrived selling profiles of specific doctors’ proscribing patterns to pharmaceutical marketing folks. Later, they moved into aggregating information on individual patients—anonymized, of course, in accordance with HIPAA rules.
Despite those rules, some are concerned about patient privacy. IMS does not disclose how it compiles their patient dossiers, and it may be possible that records could, somehow someday, become identifiable. One solution would be to allow patients to opt out of contributing their records to the collection, anonymized or not, as marketing data firm Acxiom began doing in 2013.
Of course, it isn’t quite so simple for the consumer. Each health record system makes its own decisions about data sharing, so opting out could require changing doctors. On the other hand, many of us have little choice in our insurance provider, and a lot of those firms also share patient information. Will IMS move toward transparency, or continue to keep patients in the dark about the paths of their own medical data?
Cynthia Murrell, July 22, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
There is a Louisville, Kentucky Hidden Web/Dark
Web meet up on July 26, 2016.
Information is at this link: http://bit.ly/29tVKpx.
Is IBM Vulnerable to OpenText?
July 21, 2016
I read “Hey, IBM, OpenText Is Coming for You.” The write up reports that the poobah of OpenText said that its new Magellan system is “a next generation analytics platform.” Getting from Yet another OpenText system (YOTS) to the nemesis of IBM is quite a leap.
But here’s the statement, once again from the OpenText poobah, that caught my attention:
But even more interesting than the product itself, is the bullish way in which OpenText is calling out IBM Watson. “We are going to position it directly against Watson. We’re not going to shy away from that at all,” Mark said. “We think there’s a whole class of problems that enterprises want to solve themselves and what they need is an affordable platform, one that’s open and programmable to them and accessible to them and that’s going to be Magellan. So we’re going to position ourselves and stay focused directly against Watson.”
The write up explains that OpenText Magellan is better, faster, and cheaper. I have heard that before I think. But the details are interesting.
Magellan’s software is open., Its hardware is open. Its IP is owned by the licensee. Its deployment options are “run anywhere.” It is extensible by the licensee. Its ecosystem is open. Its cost is a mere one dollar sign.
And what do you think about IBM Watson? Well, its software is closed. Its hardware is closed. Its IP ownership is not the licensee’s. Watson is extensible only by IBM Global Services. IBM’s ecosystem is closed. Best of the points, IBM’s cost is six dollar signs.
OpenText is a $2 billion a year outfit. The hapless IBM is, despite its being lost in revenue space, is about $90 billion a year.
My view is that OpenText is swinging for the marketing and conceptual fences. IBM is trying to find the secret notebook that unlocks revenues.
I would point out that Fancy Dan software designed to help executives make better decisions is plentiful. Many vendors covet this niche. There is excitement ahead. Both OpenText and IBM may find that talk about smart software flows more rapidly than sustainable revenue and healthy profits. Keep in mind the high cost of technological debt. That’s one dot point which IBM and OpenText share a common point of weakness.
Stephen E Arnold, July 21, 2106
-
- Subscribe to Beyond Search
Feature archive
News archive
-
