cool hit counter Kirin Chip AI Chief Scientist, Explains How AI Chips Make Phones Smarter_Intefrankly

Kirin Chip AI Chief Scientist, Explains How AI Chips Make Phones Smarter


Source: Huawei

Summary: On August 31, Huawei announced a new generation of top AI mobile chips - Kirin 980, becoming the world's first mobile chip with a 7nm process. Kirin 980 can do face recognition, object recognition, object detection, image segmentation, intelligent translation, etc. to achieve a qualitative leap in AI recognition.

Today, we have Dr. Xianglin Rui, Chief Scientist of Kirin Chip & Solutions AI, explaining how end-side AI chips can make phones smarter.

AI chip boosted, starting with photos

With the addition of AI, how smart are phones now taking photos?

It recognizes objects and scenes, and automatically selects imaging from "implicit" to more "explicit". The AI chip automatically selects the appropriate shooting mode to bring sharper photography and recognizes 100 images in 5 seconds.

AI chips, also known as AI accelerators or compute cards, are modules dedicated to handling a large number of computational tasks in AI applications (other non-computational tasks are still handled by the CPU), according to Rexroth.

The computing scenarios for AI chips can be divided into cloud AI and endpoint AI. The computational scenarios for deep learning can be divided into three categories, namely, training in data centers, inference in data centers, and inference in embedded devices. The first two can be summarized as applications for the cloud and the latter can be summarized as applications for the endpoint.

Mobile AI chips are able to process various AI operators at a speed of about 30x to 50x, and are able to do better an image detection, image segmentation and image semantic understanding for photo scenes, for example. In addition, the voice can be heard and understood, and the service that the customer really wants is provided based on the understood intent of the customer.

In 2017, Huawei was the first in the industry to release the Kirin 970 processor, Huawei's first artificial intelligence end-side chip, which solves the performance and power consumption problems of running AI models on the end-side from the hardware level, which gives smart terminals powerful computing capabilities.

Taking image recognition speed as an example, Kirin 970 can reach about 2005 images per minute, while it can only process 97 images per minute without NPU. Compared with the previous generation, the graphics processing performance is improved by 20% and energy efficiency is improved by 50%, which can support smooth operation of 3D large games for longer time and support new generation mobile internet experience such as AR/VR.

With the powerful mobile computing power of dual NPUs, Kirin 980 achieves 4500 images per minute, which is 120% faster than the previous generation and much higher than the same period in the industry.

Constructing a platform and playing smart combinations

Rexroth said the future is the integration of four aspects such as cloud, pipeline, AI chips and terminal capabilities. The cloud side and the end side can do better synergy, can exchange capabilities more quickly, and do coordination on capabilities. We have to have such an open platform that can interface all the capabilities within our ecosystem.

To maximize the performance of Huawei's NPU in handling AI tasks and to open up the best end-side AI capabilities to Android developers around the world to build rich AI applications, Huawei built the HiAI open platform in 2017. It has built a three-tier AI open ecology: open service capability, open application capability and open chip capability. Enables third-party developers to implement AI features on the end-side easily, in real-time, efficiently, and flexibly, thereby greatly reducing the cost of building AI features on the end-side.

HiAI open platform in natural language understanding (NLU), automatic speech recognition (ASR) and computer vision (CV) in three categories, 20 + open api and complete tool platform, to achieve out-of-the-box use, to help three-party developers quickly enrich the capabilities of their products and application scenarios, so that AI is within reach.

Currently, some partners have "enjoyed" the HiAI platform empowerment and NPU computing power to bring significant improvements in product capabilities and experience.

Microsoft Translate, for example, has seen an overall 3x speedup in offline image as well as text translation end-to-end through the acceleration of the HiAI platform.

In computer vision, the performance improvement is particularly outstanding, such as on Prisma, a photo of the style conversion of the time roughly 10 seconds, through the HiAI platform accelerated 4 times, only about 3 seconds to complete the processing.

At the same time, it can also apply HiAI open scene recognition capabilities, intelligent recommendation of the current photo suitable style filter, the user does not need to choose from dozens of hundreds of filters at every turn, to further enhance the conversion efficiency of users.

Energy efficiency ratio is also not bad, fast users live on the terminal device, more than 5 minutes, there are some terminals will heat up, the shell temperature of more than 36 degrees, the NPU energy efficiency ratio than the CPU 50 times higher, through the HiAI platform optimization, long time live without pressure.


Recommended>>
1、The new seven happy hall three male there is no auxiliary three male have hang cheating
2、Indian media ranked the best smartphone of 2017 and the number one is actually it
3、One linux command a day for the less command
4、Maersk and IBM to form joint venture MOL and Weathernews to build vessel and cargo matching platform
5、Will you be replaced by artificial intelligence in the future

    已推荐到看一看 和朋友分享想法
    最多200字,当前共 发送

    已发送

    朋友将在看一看看到

    确定
    分享你的想法...
    取消

    分享想法到看一看

    确定
    最多200字,当前共

    发送中

    网络异常,请稍后重试

    微信扫一扫
    关注该公众号