A machine-learning accelerator is resistant to the two most common types of attacks: Side-channel and bus-probing.

This accelerator can keep a user’s sensitive data private while still enabling huge AI models to run efficiently on devices. To that end, a team of researchers from MIT and the MIT-IBM Watson AI Lab developed several optimizations that enable strong security while only slightly slowing the device.

Moreover, the added security does not impact the accuracy of computations. This machine-learning accelerator could be particularly beneficial for demanding AI applications like augmented and virtual reality or autonomous driving.

While implementing the accelerator chip would make a device slightly more expensive and less energy-efficient, that is sometimes a worthwhile price to pay for security, said Maitreyi Ashok, an electrical engineering and computer science (EECS) graduate student at MIT and lead author of a paper on the subject.

Security from the Beginning
“It is important to design with security in mind from the ground up. If you are trying to add even a minimal amount of security after a system has been designed, it is prohibitively expensive. We were able to effectively balance a lot of these tradeoffs during the design phase,” Ashok said.

Schneider Bold

The researchers targeted a type of machine-learning accelerator called digital in-memory compute (IMC). A digital IMC chip performs computations inside a device’s memory, where pieces of a machine-learning model end up stored after moving over from a central server.

The entire model is too big to store on the device, but by breaking it into pieces and reusing those pieces as much as possible, IMC chips reduce the amount of data that must move back and forth.

The catch is IMC chips can be susceptible to hackers. In a side-channel attack, a hacker monitors the chip’s power consumption and uses statistical techniques to reverse-engineer data as the chip computes. In a bus-probing attack, the hacker can steal bits of the model and dataset by probing the communication between the accelerator and the off-chip memory.

Digital IMC speeds computation by performing millions of operations at once, but this complexity makes it tough to prevent attacks using traditional security measures, Ashok said.

She and her collaborators took a three-pronged approach to blocking side-channel and bus-probing attacks.

Data Split into Random Pieces
First, they employed a security measure where data in the IMC split into random pieces. For instance, a bit zero might split into three bits that still equal zero after a logical operation. The IMC never computes with all pieces in the same operation, so a side-channel attack could never reconstruct the real information.

But for this technique to work, random bits must end up added to split the data. Because digital IMC performs millions of operations at once, generating so many random bits would involve too much computing. For their chip, the researchers found a way to simplify computations, making it easier to effectively split data while eliminating the need for random bits.

Second, they prevented bus-probing attacks using a lightweight cipher that encrypts the model stored in off-chip memory. This lightweight cipher only requires simple computations. In addition, they only decrypted the pieces of the model stored on the chip when necessary.

Third, to improve security, they generated the key that decrypts the cipher directly on the chip, rather than moving it back and forth with the model. They generated this unique key from random variations in the chip introduced during manufacturing, using what is known as a physically unclonable function.

“Maybe one wire is going to be a little bit thicker than another. We can use these variations to get zeros and ones out of a circuit. For every chip, we can get a random key that should be consistent because these random properties shouldn’t change significantly over time,” Ashok said.

They reused the memory cells on the chip, leveraging the imperfections in these cells to generate the key. This requires less computation than generating a key from scratch.

Security for Machine-Learning Workloads
“As security has become a critical issue in the design of edge devices, there is a need to develop a complete system stack focusing on secure operation,” said Anantha Chandrakasan, MIT’s chief innovation and strategy officer, dean of the School of Engineering, and the Vannevar Bush Professor of EECS and senior author of the paper. “This work focuses on security for machine-learning workloads and describes a digital processor that uses cross-cutting optimization. It incorporates encrypted data access between memory and processor, approaches to preventing side-channel attacks using randomization, and exploiting variability to generate unique codes. Such designs are going to be critical in future mobile devices.”

To test their chip, the researchers took on the role of hackers and tried to steal secret information using side-channel and bus-probing attacks.

Even after making millions of attempts, they couldn’t reconstruct any real information or extract pieces of the model or dataset. The cipher also remained unbreakable. By contrast, it took only about 5,000 samples to steal information from an unprotected chip.

The addition of security did reduce the energy efficiency of the accelerator, and it also needed a larger chip area, which would make it more expensive to fabricate.

The team is planning to explore methods that could reduce the energy consumption and size of their chip in the future, which would make it easier to implement at scale.

“As it becomes too expensive, it becomes harder to convince someone that security is critical. Future work could explore these tradeoffs,” Ashok said. “Maybe we could make it a little less secure but easier to implement and less expensive.”

Pin It on Pinterest

Share This