Chip designers received an interesting challenge this month from Jeff Hawkins, the famous founder of Palm and Handspring and an expert on the human brain: If they really want to design something intelligent, they shouldn't be doing processors. They should be making memories.

At the International Solid State Circuits Conference in San Francisco, Hawkins addressed the question of "Why can't a computer be more like a brain?" He has been preoccupied with that question for decades and recently has been working on the answers for his new start-up, Numenta.

Hawkins, inventor of the Palm Pilot, wrote about his research on the brain in 2005, when he published a top-selling book, "On Intelligence." He created the Redwood Neuroscience Institute in 2002 and in 2005 co-founded Numenta with his longtime partner Donna Dubinsky. Their company is working on software that is brain-like in recognizing patterns. Hawkins says the software may have applications in computer visual recognition, among other areas.

Hawkins said computers today are fast at solving calculations but terrible at doing things that children can do, like understanding simple stories or recognizing the difference between a cat and a dog. That's because computers are designed with Von Neumann architecture, with a processor connected to memory. The brain, by contrast, has a very different design. It is a "hierarchical memory system." At the lowest level, the brain's neural connections store memories such as the shape of an object. At the next level up, the pattern associated with a human form is stored. And at the highest level, the brain recognizes the person as Bill Clinton.

The brain operates on the same formula, or algorithm, in every region, according to Hawkins. It doesn't have one algorithm for processing vision or a different one for processing sound. Each part of the brain is also self-training. You are born with nothing and learn. Each part of the brain learns based on what information is passing through it.

"This is a basic blueprint for how the memory system works," he said.

All the nodes, or neuron connections - the places where memories are stored - do the same thing. They look at patterns of data coming in. They look for common spatial patterns and sequences of those patterns, and store them. If a node learns sequences that it sees, then it passes that information upward, like passing a name upward instead of the details. You have fast-changing patterns at the bottom of the hierarchy and slow-changing patterns at the top. Words and concepts are at the top. Each node can predict what the next sequence will be and passes that along. You can generate a complex pattern.

"What this does is you build a hierarchical model of the world," he said.

The recognition process improves because the memories are essentially time stamped, where the brain stores a sequence of memories related in time. For instance, the brain might have a hard time recognizing a helicopter in a snapshot of a scene. But if the memory stored shows how that helicopter moves within the scene, it is much easier to recognize. Numenta's software can accomplish this kind of recognition task today.

When shown a crude picture of a helicopter, Numenta's software can recognize it 19 percent of the time. But when the helicopter is moving and several frames of the memory are stored together, the software can recognize the helicopter 52 percent of the time, Hawkins said.

Numenta released its software development tools about 10 months ago. Today, there are 100 to 200 developers and eight partner companies. The partners are using the software for applications such as video games, auto lane-change prediction, speaker voice identification, visual object recognition and process control of power networks.

"We have a long way to go," Hawkins said. "We're not the only ones doing this."

The goal is to create a computer that is efficient at learning to recognize things, improves itself over time, can be generally used across a lot of different industries and has an ability to make predictions.

"The goal is not to build humans, but to solve problems that are much harder" than what humans can solve, Hawkins said.

Hawkins said his software can make use of the multiple processing cores engineers are putting on each chip. But he would like to see engineers create chips with hierarchical memory systems that have their own processing capabilities. I hope they can do it.

---

(c) 2008, San Jose Mercury News (San Jose, Calif.).

Visit MercuryNews.com, the World Wide Web site of the Mercury News, at http://www.mercurynews.com.

Distributed by McClatchy-Tribune Information Services.

---

TO SUBSCRIBE TO PLUGGED IN

Items in the PLUGGED IN package are not included in your MCT News Service subscription. You can subscribe to the PLUGGED IN package or purchase the items a la carte on MCT Direct at www.mctdirect.com. To subscribe, please call Rick DeChantal at Tribune Media Services at (800) 245-6536 or rdechantaltribune.com. Outside the United States, call Tribune Media Services International at +1-213-237-7987 or e-mail tmsitribune.com.

-----

PHOTO (from MCT Photo Service, 202-383-6099).

AP-NY-02-27-08 1500EST