Google Will Improve Machine Vision

Google engineers have developed a machine vision technique able to bring high power visual recognition to ordinary desktop and even mobile computers. It is claimed that the system is able to recognize 100,000 object types within a picture in a few minutes.

The tech giant seems to have improved the fairly standard method of applying convolutional filters to a photo to pick out objects of interest. This appeared tricky, as the filters must need a sample of at least one per object type, i.e. if you are scanning Facebook for cats you will need a filter which identifies cats. In other words, the technique is limited to a small number of categories, or you will need a huge database.

A report, penned by Mark Segal, Tom Dean, Mark Ruzon, Jonathon Shlens, Sudheendra Vijayanarasimhan and Jay Yagnik, describes the method which is able to speed things up by using hashing. The matter is that locality sensitive hashing looks up the results of each step and instead of applying a mask to the pixels and summing the result, it hashes the pixels and then uses them as a lookup in a table of results. In addition, they also use a rank ordering method indicating which filter is the best match for further evaluation.

The experts point out that the change to the basic algorithm has resulted in a speed up of almost 20,000 times faster. With 100,000 object detectors that require more than a million filters to be applied to multiple resolution scalings of the target image, the set up could recognize an object in less than twenty seconds. It should be noted that the hardware used was a single multi-core machine with 20GB of RAM.

Hacking Is Not That Dangerous

It turned out that Intel subsidiary McAfee has exaggerated the cost of hacking in its fundamental study which was used to form the basis for American government cyber security policy. A study was conducted back in 2009 by the Centre for Strategic and International Studies and claimed that hacking cost the world economy $1 trillion.

Now US President keeps quoting this figure, as well as intelligence officials and members of Congress, in order to press for legislation on cybercrime protection. However, it recently turned out that $1 trillion was a huge exaggeration by McAfee. This was revealed by another study that was carried out by CSIS. The outfit discovered multiple flaws in the methodology of McAfee’s study and made a conclusion that a specific number would be much more difficult to calculate. According to the new research, the proper number could be $100 billion to $500 billion – although it’s a pretty high figure, it is hardly a trillion.

This is not the first time the McAfee’s figure was believed to be overstated – before, two principal Microsoft researchers had the same doubts. Actually, the CSIS explained that the United States might lose as little as $20 billion to $25 billion annually to cybercrime or as much as $100 to $140 billion. Both researches were underwritten by McAfee, which is considered to be one of the largest security technology vendors. In the meantime, the bias of the 2009 report was commented on at the time.

Media reports suggest that the companies can have a hard time trying to figure out what was stolen, while there are plenty of more complex economic issues out there which keep the surveys from being accurate. Now, even though the figures are 1/3 lower than the last survey, experts still have doubts that McAfee has got it right. It is clear what McAfee gets out of it if the figure is high – it is known that the company has helped the Department of Defense design a secure infrastructure and the latter has cited McAfee’s $1 trillion overestimate to argue for the expansion of cybersecurity programs.

Although hackers from abroad are definitely an increasing threat, it looks like the arguments for such expansions were based on McAfee’s earlier study results.