Amazon Plans to Design AI chips for Alexa
Alexa, an intelligent personal assistant, developed by Amazon, lets users perform various tasks such as making to-do lists, setting alarms, streaming podcasts etc. It is also capable of voice interaction and can provide information on weather or give news updates. The list of functions that Alexa can perform is enormous.
To add to this list and to improvise the functioning of Alexa, Amazon is now planning to design an AI chip for speeding Alexa’s responses and to work on hardware powered by Amazon’s Alexa. The AI chip by Amazon will be designed to help its voice-enabled products more efficiently handle tasks. The idea is to perform more data processing on the device than in the cloud. Since maximum data processing will be done on the device, Alexa would be able to parse the user’s voice pattern against commonly used phrases on the device itself, without having to bounce up to the cloud and back. This could improve Alexa’s Ability to infer and respond faster to the commands.
Another benefit of having data process on the device is that it will be able to perform certain functions without Internet connection. Also, the user’s personal information can be now saved on the device. Additionally, the process would be simplified as accessing the cloud every time will not be required.
According to the reports which first surfaced on The Information , the AI chips will also be used for Amazon Web Services (AWS), which would be geared toward machine training, to allow users a more intuitive experience. The technical specs for the chip itself weren’t detailed, but the report noted that Annapurna Labs, an Israeli chip maker has been tasked with the chip’s development.
Our take on this:
Google and Apple already have custom hardware and Amazon is beginning to develop one to stay competitive in the smart home hardware market. Alexa made performing basic tasks extremely simple and with the AI chip, it will be faster and more efficient. As proposed, most of the data processing will be performed on the device itself – even without the internet. We could expect that few years down the line, AI systems will be so small that they could even fit in the palm of your hands!
An avid reader and blogger who loves exploring the endless world of data science and artificial intelligence. Fascinated by the limitless applications of ML and AI; eager to learn and discover the depths of data science.
