10
March
2010
|
03:24 PM
America/Los_Angeles

Intel Opens Up Its Intel Labs At Berkeley - Reveals Research Projects

Intel hosted an open house this afternoon at its Intel Labs Berkeley. The lab has been in existence since November 2001.

Some of the projects revealed:

Intel Mash Maker

This is a browser extension that tries to understand the content on a web page and suggest additional pages. You can download it here: mashmaker.intel.com

[I remember Autonomy having a similar experimental product about ten years ago. It also worked with documents that you were working on. A small window at the bottom of your screen would find and display relevant information from the Internet.]


Common Sense

This is a project that uses inexpensive sensors that would be carried by people as they go about their day. It would measure and collect environmental data such as air pollution. www.communitysensing.org

[I was recently in Paris and saw a presentation that used a watch equipped with sensors that measured, pollution, and noise levels as people carried the watch around town. The information was collated on a public website.]


Confrontational Computing

This project tries to understand why people argue on the Internet and also to find information that is disputed. One current example is the information about the toxicity of vaccines. disputefinder.cs.berkeley.edu

[This is interesting because there is a tremendous amount of pseudoscience around serious medical issues. Vaccines, for example, can be a life and death decision yet there is lots of provenly bad information around.]

RouterBricks

A project that seeks to use generally available computing components to build routers rather than requiring specialized components. This would allow more flexibility in building networks, including the networks that serve the Internet. Intel says, "what the PC did for computing could be extended to network infrastructure and programming." www.routerbricks.org

[It's not surprising that Intel would be keen to see this project succeed :)]

Power Aware Perception

Intel says that in the future our mobile phones will know our social networks, our favorite places, what we do, etc. It will even teach us how to do things. It will become "your best friend and personal factotum."

It is working on developing 'perception algorithms.' But these are sophisticated computing problems and Intel's power management systems have been developed to adjust application performance on-the-fly, so that mobile phones can handle the compute intensive tasks related to 'perception.'

[I would think that a lot of the computer intensive tasks would be better run in the cloud. And that would be the same for any other computer intensive tasks. There's not much sense in running those applications on your mobile phone. But maybe I'm missing something here.]

Yada

This is a project that seeks to ease the development of software that runs in parallel. Today's microprocessors have multiple processor cores. To take advantage of this architecture, software applications need to be written so that they can be executed in parallel.

Yada is a programming language that looks like a sequential programming language but it detects which parts of the program can be executed in parallel.

[This is very important. It's very difficult to program for parallel execution, and tools such as Yada, are necessary so that the advances in multi-processor core multiprocessors can be used to their full capacity. Just because the latest microprocessor has multiple cores doesn't mean that your applications can take advantage of it. It's one of the dirty little secrets of multi-core multiprocessors.]

Data Mining for Anomaly Detection

Malicious software and botnets operated by hackers, are a big threat to the security of the Internet and to individual companies. The recent attack on Google, and about 100 other large companies, was achieved through the use of botnets.

This project analyzes the behavior of individual computing devices on the network and also servers. It looks for anomalies in their behavior that could signal that they have become recruited into a botnet. The project also looks for "data poisoning" where data has been deliberately inserted into company databases for nefarious purposes.

[Botnets and malicious software are a huge problem. It's further compounded by the fact that there are lots of legitimate software applications, such as cookies, that can be mistaken for malware. This is an area where there is a war of epic proportions going on between the malware authors, and the companies that create defensive software. It's a problem that won't be solved anytime soon.]

You can see some of the other research projects here:

Intel Labs Berkeley - Projects

- - -

[Please note: I am a member of the Intel Insiders - a small group of advisors to Intel.]