Friday, March 30, 2012

Data growth and standards

Data growth and standards:
An exploration of relevant open standards
Peter Haggar, Senior Technical Staff Member, IBM
Summary:  Examine the challenges presented by the explosion of data, the analytics thereof, and an introduction to some standards relevant to these challenges. A sample scenario depicts a system where large amounts of data are ingested, understood, and manipulated, and where some specific standards promote integration and interoperability.

'via Blog this'

Tuesday, March 27, 2012

Improve Your VDI Performance With New Add-in Hardware From Teradici

Improve Your VDI Performance With New Add-in Hardware From Teradici:
teradici.jpgRemember the days when you needed to add a graphics card in your PC to support higher resolution displays? That is the idea behind a new add-in card from Teradici that is used to boost the performance of VDI implementations. One of the issues with VDI is that having multiple virtual machines come online at the same time (such as around 9 am, when everyone is coming to work and turning on their computers) can bog down a server. Not any longer.
Sponsor

The card, called the APEX 2800 and pictured above, plugs into any PCIe slot and runs some software drivers to talk to VMware vSphere 4.1, ESXi 5.0 and View 5.0. By offloading image encoding to a separate hardware encoding card, as much as 50% of, CPU capacity can be recovered, at least according to Teradici's own tests. This means you can have each hypervisor serve up more virtual desktops. If you are running a lot of graphics-intensive VMs, such as Windows 7 with the aero interface enabled, you might want to look into this option. The card costs $2000 and is available from both HP and Dell online stores. The server offload card is compatible with all existing PCoIP clients and VMware View software clients.
Discuss

I'll Believe It When I'm Shooting Them Out Of The Sky: Iffy Autonomous Quadrocopter Taco Delivery Service

I'll Believe It When I'm Shooting Them Out Of The Sky: Iffy Autonomous Quadrocopter Taco Delivery Service: taco-copter.jpg

Those aren't f***ing tacos -- that's a clown nose, dummy!

Allegedly (and I stress allegedly HARD and in a real husky voice that makes you weak in the knees and pee a little), a startup in San Francisco wants to sell Mexican food and make deliveries with the use of an autonomous quadrocopter. Obviously, it'll never happen because I'll lure all the unsuspecting taco-copters into a warehouse and then beat them with a stick. It's too easy! PLUS BURRITOS. But seriously, the logistics behind an operation like this sound like a nightmare. People stealing your copters, copters crashing into people and causing accidents -- the possibilities are endless. That said, if the service actually does get off the ground, I'll eat my words. Jk jk -- NACHOS. "Hey GW -- that's nat-cho cheese!" Ahahahahaha -- they're not my chips either but f*** my roommate!

Official Site

Thanks to the TOASTer, Naterade and MacLean/Renegade, who get their tacos delivered the old fashioned way: on the back of a donkey.

Leaked AMD 7990 Dual-GPU Specs Leak, Try To Steal Thunder From Nvidia’s Party | Gadgetsteria

Leaked AMD 7990 Dual-GPU Specs Leak, Try To Steal Thunder From Nvidia’s Party | Gadgetsteria: "April "

'via Blog this'

Comic for March 27, 2012

Comic for March 27, 2012:

Thursday, March 22, 2012

Secret Nuclear Drone Plan Nixed by 'Political Realities'

Secret Nuclear Drone Plan Nixed by 'Political Realities': A government lab hatched a secret plan for nuclear-powered drones, a new document shows. But the project had to be abandoned for "political" reasons.

Wednesday, March 21, 2012

The beauty of machine learning? It never stops learning

The beauty of machine learning? It never stops learning:
George Gilbert of GigaOM Pro, Currie Boyle of IBM, Alexander Gray of Skytree, Mok Oh of PayPal, and Amarnath Thombre of Match.com at Structure:Data 2012
(c) 2012 Pinar Ozger. pinar@pinarozger.com
Business not using machine learning to augment the products and services will find it difficult to compete in the future according to panelists at GigaOm’s Structure:Data event on Wednesday. Not only will these companies be at a competitive disadvantage at first, it will get worse. Why? Machine learning solutions will only gain more intelligence with additional data and techniques.
Currie Boyle, a Distinguished Engineer at IBM, explained how natural language processing and continual machine learning transforms online sellers, allowing them to interact with its customers. “These products help with guided selling on the web across online retail sites. More importantly, they try to understand the unsuccessful transactions to improve machine learning. That can transform clients from being relatively static to human-like; the more it’s used, the more successful for you and others.”
The beauty is that such machine learning solutions can help both buyers and sellers in a transaction in ways that weren’t possible just a decade ago. Mok Oh, Chief Scientist at PayPal noted this: “Human loops trying to match buyers and sellers is inefficient. Machine learning makes it scalable and cost-effective to connect them in a world of unstructured data across millions of sites and products.”
While this machine learning sounds like a magical solution, it’s not bulletproof because it depends on human observation that varies from inputs. Amarnath Thombre, SVP, Strategy and Analytics at Match.com believes strongly that successful businesses will be those that observe their customers. But the company found a problem: Match.com customer profiles sometimes differed from how those customers behaved in real life.
“There’s no one single formula to match people, but we had millions of data points for successful matches,” Thombre said. “Mining it provided us an equation for success but that wasn’t enough. We saw a 2x increase in success after modifying our machine learning algorithms to account for the differences that customers had in their profile vs real life.”
Although machine learning is powering success, not all business are using it. That’s a problem says Alexander Gray, CTO, Skytree, suggesting that in the future every business will be a data-driven enterprise. “Every organization will need to do machine learning to augment human decisions” As data grows and machine learning improves, however, that will require more computing horsepower and multiple iterations of data interpretation.
Curry summed it up best by saying that this constant learning can help retain customers, even as barriers for customer departures are decreasing. ”Models change every quarter so machine learning has to continue learning. We need to keep getting the right data sets and asking the right questions for machine learning to help customer retention.”
Watch the livestream of the event here.
Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.






Supreme Court Rejects Business Patents On Abstract Ideas From Nature

Supreme Court Rejects Business Patents On Abstract Ideas From Nature: The Supreme Court ruled against Prometheus Laboratories, saying that an abstract idea based on natural phenomena was not eligible to be patented.

Army's Self-Driving Trucks Let Humans Watch for Bombs | Danger Room | Wired.com

Army's Self-Driving Trucks Let Humans Watch for Bombs | Danger Room | Wired.com:

'via Blog this'

Tuesday, March 20, 2012

A Computer Program to Model Human Decision-Making


via Mashable! by Sam Laird on 3/20/12


Cheating at chess? A computer program that can call you out may be here soon. A computer science professor and chess master has spent the past five years at work on a system to determine when players get unfair electronic assistance with their kings and queens.
Kenneth Regan’s system essentially works as a mathematical proof using the moves that would be made by a computer chess program to learn more about human decisions. Some say his work could have powerful implications beyond the chessboard as well.
“What he is doing, what these people are doing, is they are trying to model how people make decisions,” Dr. Jonathan Schaeffer tells The New York Times.
A model of human-decision making would have tremendous value for sites and services that want a clearer window into how humans think. Online retailers could benefit especially.
Check out the video above for more information on on what Dr. Regan is up to.
Thumbnail image courtesy of Flickr, LdF
More About: artificial intelligence, e-commerce
For more Tech coverage:

Monday, March 19, 2012

How scientists can manipulate us with games to gather data

How scientists can manipulate us with games to gather data:

Smartphones have turned us into an army of accidental data collectors, checking in at locations, taking photos, recording audio, even gauging network speeds wherever we go. This rich data can be used for research, but because people are creatures of habit, there are sometimes gaping blind spots.
“Flickr has thousands of photos of the front of the Lincoln Memorial. But who takes a picture of the back? Very few people,” said Fabian Bustamante, associate professor at Northwestern University, in a statement. He proposes “soft controlling” people into gathering information needed for research by using games or social networking apps. The apps would nudge people into completing a task at a specific location, then reward them with points or incentives.
Bustamante and his group at Northwestern University whipped up a location-based, augmented-reality Android game called Ghost Hunter to test the theory with students. In the game, a player wanders around her neighborhood looking for ghosts and other monsters, and when the ghouls are in range, the app switches to a live view with an overlay of cross-hairs. The player shoots the ghost and a photo of the location is uploaded.
Just playing the game was enough of an incentive to get participants to wander off their beaten-track, according to graduate student John P. Rula. Rula was the lead author of a paper on the project.
Full sets of crowdsourced photos can be used to do neat things like create 3D models of landmarks or buildings. Another possible use for this technology is tracking noise pollution. Mobile phones and tablets already have built-in microphones that can be used to measure ambient noise, but what would make you whip out your iPhone and hit a button right as the N train rumbles into the station? How about a reward such as the chance to win $5 at Starbucks, or something less tangible like killing a brains-hungry zombie for points?
There are already plenty of apps that reward you for “checking in” at a location, including Foursquare and Gowalla. Soft control functions could be easily integrated into existing apps with about 25 lines of additional code, according to the paper.
Researchers are quick to point out the idea isn’t to trick people into sharing information; all participants would be notified and their data anonymized when uploaded.
The idea may sound a bit odd at first — who would agree to be “soft controlled”? But many people are actually eager to become citizen scientists, as evidenced by the popularity of the SetiQuest Explorer app, which allows anyone to help monitor radio signals in space to help search for extraterrestrial life.

Filed under: mobile, offBeat, VentureBeat

Friday, March 16, 2012

ATLAS project

ATLAS project:

'via Blog this'

Robots Using ROS: Marvin autonomous car (Austin Robot Technology/UT Austin) - ROS robotics news

Robots Using ROS: Marvin autonomous car (Austin Robot Technology/UT Austin) - ROS robotics news: " the video above, these intersections for autonomous vehicles can handle far more vehicles than intersections designed for human-driven vehicles. For more information, you can watch a longer clip and read Kurt Dresner and Peter Stone's paper, "A Multiagent Approach to Autonomous Intersection Management"

Many people have contributed to the development of Marvin in the past. Current software development, including porting to ROS, is being led by Jack O'Quin and Dr. Michael Quinlan under the supervision of Professor Peter Stone."

'via Blog this'

Google's Flight Search engine adds 500 international destinations, more on the way

Google's Flight Search engine adds 500 international destinations, more on the way:
Google's Flight Search -- which was conjured up after its acquisition of ITA -- has been giving Kayak and Bing Travel a run for their respective dollars here in the States, but an airline search engine is only as good as its reach. This week, the company announced that over 500 airports outside of the US are now being included in the results (but, sadly, only if you start in the US), and while no specific promises were given, Google says that it's "working hard on expanding [its] global coverage and adding more routes in the future." Oh, and for those looking to disconnect in Samoa, APW totally made the cut here. Happy trails!

[Thanks, sciwiz]
Google's Flight Search engine adds 500 international destinations, more on the way originally appeared on Engadget on Fri, 16 Mar 2012 10:47:00 EDT. Please see our terms for use of feeds.
Permalink Jeff Huber (Google+)  |  sourceThe Official Google Search Blog, Google Flights  | Email this | Comments

Friday, March 9, 2012

JOBS Act passes: What the new crowdfunding bill means for startups

JOBS Act passes: What the new crowdfunding bill means for startups:
The JOBS ((Jumpstart Our Business Startups) Act that passed in the House today contains some big changes for crowdfunding startups. It now moves on to the Senate.
Before today, it was illegal for a startup to solicit investors on platforms like Twitter or Kickstarter. For startups raising $1 million or less, anyone can now buy up to $10,000 or 10 percent of the annual income (whichever is less) in equity. (Anyone who wants to get in on the ground floor of my bonsai pets business, send your checks to WeWork Midtown, care of VentureBeat.)
One of the principle drivers behind the IPO filings of Facebook and Zynga was the 500-shareholder rule, a vestige of the Securities and Exchange Act of 1934, which said that any company with more than 500 shareholders has to open its financials to the SEC like a public company. But under the JOBS Act, anyone who gives $10,000 or less will not count toward this limit. The act also raises the shareholder limit from 500 to 1,000.
Startups can opt to raise as much as $2 million in this manner; but if they go the crowdfunding route, they will have to provide audited financial statements to their investors. And while raising capital from the crowd is pretty nifty on Kickstarter, it has some drawbacks for startups. Here are some highlights from our FAQ on crowdfunding:
First, startups must understand that minority stockholders have certain significant rights under state law, including voting rights, the right to inspect the company’s books and records, the right to bring a derivative claim on behalf of the company, and certain protections against oppression by the controlling stockholders. Indeed, the more stockholders a startup has, the greater the likelihood that a disgruntled stockholder will cause problems, including filing lawsuits.
Second, having hundreds of stockholders is an administrative nightmare and will be time consuming and costly. Presumably, each stockholder will be required to execute a subscription agreement and/or stockholders’ agreement to address key issues such as transfer restrictions, rights of first refusal, and drag-along rights. There will also be administrative issues relating to voting and stock transfer issues.
Third, startups will likely have difficulty raising funds from VCs and other sophisticated investors if they have hundreds of unsophisticated stockholders. Needless to say, few sophisticated investors will want to sit on the board of directors of such a company due to the risks of lawsuits relating to director liability, and I would assume D&O liability insurance rates will skyrocket for these companies.
The JOBS Act also makes it easier for small companies to go public by increasing the offering threshold for companies exemepted from SEC regulation from $5 to $50 million on companies. Additional regulations will be phased in over a five-year period for companies that stay under $1 billion in revenue.
This is a huge win for companies like SecondMarket. Companies can now stay private with twice as many investors and go public with a much smaller revenue stream. That will mean big growth in the small and medium cap IPOs that SecondMarket is counting on to replace the revenue it is losing when Facebook goes public.

Filed under: deals, Entrepreneur, VentureBeat