Oct. 30 (Portaltic/EP) –
OPPO has announced the implementation of the Mixture of Experts (MoE) architecture on your deviceswith which he ensures that will improve the efficiency of Artificial Intelligence (AI) processing, in addition to laying the foundation for “future innovations in the integration of AI with mobile hardware.”
The MoE architecture is a technology capable of dividing an AI model into small neural networks so that they act separately as if they were different experts. In this way, the networks are activated selectively depending on the type of information introduced, to perform a joint task more effectively.
In this sense, OPPO has announced that it has started implementing the MoE architecture on their mobile devicesin order to improve the efficiency of AI processing, to reduce the computational costs of this technology.
As the company explained, as AI technology evolves and offers more possibilities for users, the tasks carried out on the device also increase. However, it must be taken into account that the Large AI models demand “substantial” computing power. This can affect device performanceespecially if they have limited hardware resources.
In order to overcome these limitations, OPPO has Collaborated with “major chipset suppliers” for implement MoE on their smartphones. It is also the first technology that does so, as highlighted in a statement, and with this, lays the foundations for “future innovations” in the integration of AI with mobile hardware.
Specifically, this architecture manages to improve AI processing efficiency, in addition to reduce computing and data transfer consumption. According to laboratory tests conducted by OPPO, the MoE architecture achieves accelerate the speed of AI tasks by approximately 40 percent.
This is because the devices can carry out tasks with lower demand for resources and improving energy efficiency. Likewise, OPPO has stressed that, as a result, the architecture allows it to offer faster AI responses, longer battery life and greater privacy“as more tasks are handled on the terminals.”
Following this line, the company has also clarified that, at reduce the computational costs of AI, MoE makes it easier for more affordable smartphone models to also perform complex AI tasks.
“The on-device MoE architecture opens new opportunities for the industry to make “Advanced AI capabilities are more accessible to a broader audience.”OPPO has declared.
With all this, the technology company has expressed its commitment to the advancement of AI and making it available to more users. In fact, he recalled that has more than 5,860 patent applications in the field of AI and that it will continue to invest in R&D to advance in this area. He has also made reference to his AI Center, which has stood out as “key” piece to consolidate your research efforts of AI.
Add Comment