Making Speedy Memory Chips Reliable

IBM believes a new way of encoding the bits in phase-change memory will make it reliable enough for use in servers.





Long-term memory: Each cell in this 200,000-cell phase-change memory chip can store multiple bits of data reliably over a period of several months. 

IBM researchers have developed a programming trick that makes it possible to more reliably store large amounts of data using a promising new technology called phase-change memory. The company hopes to start integrating this storage technology into commercial products, such as servers that process data for the cloud, in about five years.

Like flash memory, commonly found in cell phones, phase-change memory is nonvolatile. That means it doesn't require any power to store the data. And it can be accessed rapidly for fast boot-ups in computers and more efficient operation in general. Phase-change memory has a speed advantage over flash, and Micron and Samsung are about to bring out products that will compete with flash in some mobile applications.

These initial products will use memory cells that store one bit each. But for phase-change memory to be cost-competitive for broader applications, it will need to achieve higher density, storing multiple bits  per cell. Greater density is necessary for IBM to achieve its goal of developing  phase-change memory  for high-performance systems such as servers that process and store Internet data much faster.

The IBM work announced today offers a solution. In the past, researchers haven't been able to make a device that uses multiple bits per cell that works reliably over months and years. That's because of the properties of the phase-change materials used to store the data. Scientists at IBM Research in Zurich have developed a software trick that allows them to compensate for this.

Each cell in these data-storage arrays is made up of a small spot of phase-change materials sandwiched between two electrodes. By applying a voltage across the electrodes, the material can be switched to any number of states along a continuum from totally unstructured to highly crystalline. The memory is read out by using another electrical pulse to measure the resistance of the material, which is much lower in the crystalline state.

To make multibit memory cells, the IBM group picked four different levels of electrical resistance. The trouble is that over time, the electrons in the phase-change cells tend to drift around, and the resistance changes, corrupting the data. The IBM group has shown that they can encode the data in such a way that when it's read out, they can correct for drift-based errors and get the right data.

The IBM group has shown that error-correcting code can be used to reliably read out data from a 200,000-cell phase-change memory array after a period of six months. "That's not gigabits, like flash, but it's impressive," says Eric Pop, professor of electrical engineering and computer sciences at the University of Illinois at Urbana-Champaign. "They're using a clever encoding scheme that seems to prolong the life and reliability of phase-change memory."

For commercial products, that reliability timescale needs to come up to 10 years, says Victor Zhirnov, director of special projects at the Semiconductor Research Corporation. IBM says it can get there. "Electrical drift in these materials is mostly problematic in the first microseconds and minutes after programming," says Harris Pozidis, manager of memory and probe technologies at IBM Research in Zurich. The problem of drift can be statistically accounted for in the IBM coding scheme over whatever timeframe is necessary, says Pozidis, because it occurs at a known rate.

But phase-change memory won't be broadly adapted until power consumption can be checked, says Zhirnov. It still takes much too much energy to flip the bits in these arrays. That's due to the way the electrodes are designed, and many researchers are working on the problem. This spring, Pop's group at the University of Illinois demonstrated storage arrays that use carbon nanotubes to encode phase-change memory cells with 100 times less power.

When You're Always a Familiar Face

A startup wants any Web page or mobile app to recognize faces, and claims users are becoming less sensitive about the technology.

Emotion decoded: Face.com enables any website or app to recognize faces and some simple expressions. Credit: Face.com

A startup has built technology that could give any website or app the ability to recognize peoples' faces, and even to identify their facial expression. While some see the technology as creepy, Face.com, the company behind it, argues that most users don't mind being recognized automatically online.

As computer vision software gets better, facial recognition software is becoming more common. With millions of images uploaded to services like Facebook every day, this technology can make social services more convenient, and could lead to completely new kinds of services and apps—but it also has obvious privacy implications.

For over a year, Face.com has made its technology available to software developers, enabling them to build it into a website or Web-connected apps. A website or app sends photos, which may be uploaded by users, to Face.com's servers for processing and receives details that include the location of any faces, their gender, and whether they match other photos stored by Face.com. Last week, the service was upgraded to allow it to gauge a person's mood, classifying them as happy, sad, surprised, angry, or neutral. It could already spot smiles, but has now gained the ability to classify whether a person's lips are sealed, parted, or making a kiss. These new features could perhaps be used to automatically add more detailed tags to images or to challenge people to convey a certain mood with their expression.

Within three days of the launch of these new features, one website had has begun using Face.com's mood recognition feature. Moodbattle is a website that asks visitors with webcams to compete to pull the most extreme expressions associated with particular emotions (browse the results here).

"Last month, we processed more than two billion different photos," says Face.com's CEO Gil Hirsch, who adds that usage is growing. Some 20,000 developers have signed up to tap into Face.com's technology; they can process 5,000 photos per hour for free, or pay for the ability to process more.

Hirsch says the four-year-old company recently became profitable, but he acknowledges that facial recognition still raises privacy concerns with many.

Last month, Facebook was forced to apologize to users after introducing a feature that uses facial recognition to suggest which of your friends appear in a photo, to speed up the process of tagging them. This is nearly identical to a Facebook app previously released by Face.com, although Hirsch declined to comment when asked if his company had supplied the  technology behind Facebook's facial recognition.

Google also has sophisticated facial recognition software, but it has been careful to say it does not use it in the mobile app Google Goggles, which identifies objects snapped with a phone's camera, or the recently launched Search by Image service.

Kelly Gates, a professor at University of California, San Diego, whose recent bookexamines how facial recognition technology is being developed, adopted, and understood, says concerns over the technology are largely due to an association with security and surveillance. Combined with the fact the technology interfaces with a very personal part of the body, "that means it is easy to write scary headlines about," says Gates.

However facial recognition has clear benefits in a world of online socialising. "It seems that there is a need for technology that can help with that," Gates says. "Now we are habituated to things like photo tagging it starts to seem attractive to automate it." Gates thinks the usefulness of services like Face.com's will be enough for most people to eventually accept facial recognition, just as they've accepted other technologies that initially caused privacy fears.

Hirsch argues that there has already been a shift in attitudes toward facial recognition. The technology has been accepted in desktop software for organizing photos from both Apple and Google, he points out. The key is to make sure that people feel in control of the technology, he says. For example, Face.com's service can connect with Facebook to find your friends in a photo, but it will ask permission to access Facebook data and abides by a user's privacy settings. As the technology appears in more places, people will warm to the notion of services that recognize them and their friends, says Hirsch.

Gates says that technology that recognizes moods or expressions is much less mature -- and less accurate -- than that which recognizes faces and applications for it are still unclear. Face.com's mood recognition technology likely works best with posed expressions rather than natural ones, she adds, but that could still be useful: "It's like emoticons, in that these are very simplistic expressions of emotions, but they can serve a purpose to communicate something."