Google LLC today introduced Two open-source technologies designed to make it easier for businesses to process user data in a way that meets their privacy requirements.
The first technique is a tool called Magritte for blurring objects such as license plates in videos. Another is the new version of his FHE C++ Transpiler, a privacy tool Google introduced last year. This allows applications to process encrypted datasets without first decrypting them.
Magritte is based on one of the search giant’s internal software projects. It uses artificial intelligence to automatically detect when objects containing sensitive data, such as license plates, appear in the video. Magritte then blurs the object so that the video editing team does not have to perform the task manually.
Some of our AI capabilities are powered by other open source Google tools. media pipeThe latter tools allow developers to build AI applications that can run on devices with limited computing power, such as smartphones.
Miguel Guevara, Product Manager, Google’s Privacy and Data Protection Office, said: blog post“By using this open-source code, videographers can save time blurring objects from their videos, but we know that the underlying ML algorithms can perform highly accurate detection across videos.”
Google debuted Magritte today with a new version of the FHE C++ Transpiler, an open source tool first released last June. This tool allows developers to easily implement an encryption technology called fully homomorphic encryption (FHE). This technology has received a lot of interest from researchers in recent years because it could be used to make enterprise applications more secure.
Enterprise applications store sensitive data in encrypted form to reduce the risk of cyberattacks. However, whenever we need to use the data, we have to decrypt it. Decrypted files are more susceptible to cyber-attacks as hackers can easily access their contents in the event of a breach.
The FHE encryption method that Google’s FHE C++ Transpiler tool is based on eliminates the need to decrypt data before processing it. As a result, this method helps businesses reduce the risk of cyberattacks.
In practice, however, using FHE to improve security is difficult due to several technical obstacles. One of the stumbling blocks is the prohibitive amount of infrastructure currently required to run FHE software. Another challenge is that this technology is difficult for developers to implement.
According to Google, their open-source FHE C++ Transpiler tool simplifies the task of deploying FHE. The tool can analyze code originally written to handle decrypted data and automatically adapt it to run on FHE-encrypted data. As a result, developers can create applications that can handle encrypted data with less effort than otherwise.
The new version of the FHE C++ Transpiler that Google detailed today introduces several performance optimizations. Optimizations were implemented in the circuitry used by the tool to perform its processing. In computer science, the term circuit does not refer to electronic components, but to a specialized set of computational operations performed one after the other.
Google engineers have reduced the size of the circuitry used by FHE Transpiler to process data by 50%. The result, according to the search giant, is a significant performance boost. Less infrastructure is required to run the tool, and calculations can now be performed faster.
The sheer amount of infrastructure currently required to run FHE software is one of the main reasons why this technology has not yet been widely adopted by enterprises. By reducing hardware requirements, Google’s FHE tools may make it easier for organizations to adopt his FHE.
