Protect data with efficient smartphone chip

Translate from : Beskyt data med effektiv smartphone-chip
Optimize your health with our app! We use advanced machine learning to manage chronic diseases and fitness goals. Your data is secure and efficient, even on your smartphone.

Health monitoring apps can help people manage chronic diseases or stay on track with exercise goals, just by using a smartphone. However, these apps can be slow and energy-inefficient because the large machine-learning models that drive them must be moved between a smartphone and a central memory server.

Engineers often speed up the process by using hardware that reduces the need to move so much data back and forth. While these machine-learning accelerators can streamline computation, they are susceptible to attackers who can steal confidential information. To reduce this vulnerability, researchers from MIT and the MIT-IBM Watson AI Lab have created a machine-learning accelerator that is resistant to the two most common types of attacks.

Their chip can keep a user's health records, financial information, or other sensitive data private while still enabling massive AI models to run efficiently on devices. The team developed several optimizations that enable strong security while making the device run only slightly slower.

Furthermore, the extra security does not affect the accuracy of the calculations. This machine-learning accelerator can be particularly beneficial for demanding AI applications such as augmented and virtual reality or autonomous driving. Although implementing the chip would make a device slightly more expensive and less energy efficient, sometimes that's a price worth paying for security, said lead author Maitreyi Ashok, a PhD student in electrical engineering and computer science at MIT.

"If you try to add even a minimal amount of security to a system after it's designed, it's prohibitively expensive. We were able to effectively balance many of these trade-offs during the design phase," says Ashok.

Her co-authors include Saurav Maji, a PhD student; Xin Zhang and John Cohn of the MIT-IBM Watson AI Lab; and senior author Anantha Chandrakasan, MIT's head of innovation and strategy, dean of the engineering school and Vannevar Bush Professor of Electrical Engineering and Computer Science. The research will be presented at the IEEE Custom Integrated Circuits Conference.

To test their chip, the researchers assumed the role of hackers and attempted to steal secret information using side-channel and bus eavesdropping attacks.

Even after doing millions of trials, they could not reconstruct any real information or extract parts of the model or data set. The encryption also remained unbreakable. By comparison, it only took about 5,000 attempts to steal information from an unprotected chip.

However, the addition of security reduced the energy efficiency of the accelerator chip, and it also required a larger chip area, which would make it more expensive to produce.

The team plans to explore methods that could reduce the power consumption and size of their chip in the future, which would make it easier to implement on a large scale.

Our Partners