As machine learning and artificial intelligence continue to grow as the next trends for smart devices, Google is continuing to position itself to have relevant features to engage these new technologies. One of them is TensorFlow, an open-source machine learning system that can take advantage of the new neural hardware available in devices and in the cloud.
To make TensorFlow more accessible to mobile devices, Google is now releasing TensorFlow Lite – a lightweight machine learning solution specifically designed for the hardware available in smartphones. TensorFlow Lite supports Android’s Neural Networks API, and is lightweight enough that it will still run even if only powered by device’s processor – if a dedicated neural processor (like the one in the Huawei Mate 10) is absent.
Current executions range from image recognition – using your camera to identify objects – to “smart replies”, which give users access to one touch text message replies to common queries and questions you receive over messaging.
As hardware for smartphones and mobile devices get better and better, TensorFlow Lite should start becoming a common technology for manufacturers to harness the power of machine learning in their products. Look to 2018 to see more mature executions of this current technology.
November 15, 2017 at 05:06AM