MediaTek has now provided more information about the strategy for enabling A.I. edge computing on its NeuroPilot A.I. platform, which is being showcased at CES 2018. The enhanced platform is intended to address key challenges presented by rapid growth in demand for interconnected consumer and enterprise devices. Specifically, the company hopes to bring its strategy to bear across its entire technology portfolio to address increasing demands for processing power, efficiency, battery life, and cost-effectiveness that have been brought on by that demand - and which have created a need for greater A.I. processing at the edge of networks rather than on the cloud.
The company lists a total of 5 areas on which it plans to focus the efforts of its NeuroPilot A.I. Platform and where it says the platform is already showing promise. First, MediaTek claims it can bring A.I. down to the chipset level, where it can act as what the company calls an "edge A.I. enabler," while striking a balance between energy efficiency and performance. That balance allows improved A.I. applications and features, as well as allowing them to run with more practical efficacy. At the same time, MediaTek says its new platform supports and works well with mainstream A.I. frameworks SDKs, as well as supporting popular operating systems such as Android and Linux. Finally, MediaTek plans to create its own A.I.-specific SDK for MediaTek chipsets in order to allow developers access to SoC-level functions.
Essentially, MediaTek is moving to put its extensive knowledge in the creation of power-efficient, affordable chipsets to use against the challenges of making A.I. affordable and accessible while maintaining the processing power required by current and future A.I. applications. Moreover, the company hopes to place itself in the market as an end-to-end provider of A.I.-related products and services through its NeuroPilot A.I. Platform. The news from CES 2018 about MediaTek's A.I. ambitions isn't new, as the company recently announced a shift toward A.I. processing and away from high-end mobile SoCs. However, it does provide some extra information about the company's efforts and where it hopes to go with A.I. The end goal, as the company puts it, is to create a means by which A.I. technologies can be "democratized" so that as many end-users as possible will be able to make use of the technology - in spite of just how quickly that particular portion of the industry is advancing.