China's Race for Artificial Intelligence (AI) Technology

[Stranagor]

China's Canaan to Release Its Second AI Chip This Year

PENG HAIBIN
DATE : JAN 15 2020/SOURCE : YICAI

China's Canaan to Release Its Second AI Chip This Year

,(Yicai Global) Jan. 15 -- Canaan Creative, the world's second-largest maker of cryptocurrency mining hardware, will launch its second-generation artificial intelligence chip later this year as the Chinese firm diversifies into supercomputing.

With more than 90 percent of its earnings coming from sales of Bitcoin mining machines, the Beijing-based company's financial results and stock price are closely linked to the price of Bitcoin.

But with only a limited number of the coins available, Canaan is keen to change that and distance itself from the cryptocurrency, founder and Chief Executive Zhang Nangeng told Yicai Global at the Consumer Electronics Show in Las Vegas earlier this month.

Hence the push into AI chips. These are chips that are technologically advanced enough to perform machine learning tasks, eliminating the need for human control.

Canaan's first-generation AI chip, the Kendryte K210, was released in September 2018. After overcoming a few technical setbacks, it is now being used in such application scenarios as access control, smart zones, smart home appliances, smart energy consumption and smart agriculture.

There is still room for improvement, Zhang said. Many customers don't have the expertise to directly use AI chips, so the company needs to provide overall solutions such as algorithm optimization and product module implementation. So far, just 40 to 50 staff out of 300 are involved in its chip business.

Zhang expects revenue from AI chips to match the company's conventional mining hardware unit in about two years' time. Canaan's first half revenue last year came to CNY280 million (USD40.6 million).

Canaan listed on the Nasdaq stock exchange last November, raising USD90 million and becoming the first crypto mining machine maker to go public. Shares of the company have fallen 32 percent to a closing price of USD6.14 each yesterday from an initial offering price of USD9.

https://yicaiglobal.com/news/china-canaan-to-release-its-second-ai-chip-this-year

 

[k s ahmed]

Very good progress. But sadly this might only be used to suppress the other humans.

They never caught a theif ever using cameras.

 

[JSCh]

Nature
Peng Yao, Huaqiang Wu, Bin Gao, Jianshi Tang, Qingtian Zhang, Wenqiang Zhang, He Qian (Tsinghua University), J. Joshua Yang (University of Massachusetts, Amherst)
Tsinghua University​

Novel memristor-enabled computation in memory architecture could revolutionize artificial-intelligence hardware
Conventional computing hardware are inefficient to tackle data-intensive artificial-intelligence tasks due to the underlying von Neumann architecture restriction.

23 hours ago by Tsinghua University

Artificial intelligence (AI) is revived these years because of the development of deep learning algorithms, which perform excellently in computer vision tasks (e.g. image recognition, detection) and nature language processing (e.g. machine translation, text generation). This is promising to revolutionize our society to enter an intelligent era. However, the fundamental computing hardware still face severe efficiency-issue when tackling these AI tasks, due to the limitations from the underlying von Neumann architecture. The frequent data shuffling between the separated computing and memory units accounts for large latency and power consumption in dealing with the data-intensive algorithms, which seriously limits the practical applications.

Memristor-enabled computation in memory (CIM) architecture is considered as a promising approach to substantially address the von Neumann bottleneck. Memristor is a kind of electrically synaptic device whose conductance could be easily modulated by applying appropriate voltages between the top and bottom electrodes. Organizing memristors in crossbar array and inputting the encoded voltage signals, the multiply-accumulate (MAC) computing, which is the key operation in deep neural networks, could be executed naturally in a physical manner (owing to basic Ohm’s law and Kirchhoff’s current law). Moreover, the massive MAC operations could be conducted in a parallel way within the memristor crossbars, where the calculating happens at the data-storage location. This emerging hardware with CIM architecture could strongly boost the computing efficiency in terms of the deep learning tasks and offer solid support for the wide AI scenarios from cloud to edge.

The schematic of multiple memristor-crossbar chips working jointly. The compact memristor crossbar arrays could realize parallel computation in memory naturally to go beyond von Neumann architecture. Credit: Huaqiang Wu’s Research Group

Recently, Prof. Huaqiang Wu’s research team from Tsinghua University reported the up-to-date breakthrough regarding the CIM in Nature journal, titled as “Fully hardware-implemented memristor convolutional neural network” (authored by P. Yao et al). In this work, hybrid training and spatial parallel computing techniques are proposed and demonstrated in a fully hardware-implemented CIM system to efficiently realize a convolutional neural network (CNN). The CIM system could beat its counterparts by achieving an energy efficiency more than two orders of magnitude greater than that of graphics-processing units.

Professor Huaqiang Wu, the corresponding author of this paper, comments: “Memristor device is capable to be scaled down to 2nm size. With the help of 3D process, we could further realize an incredible device integrating density in a chip as the synapses in the brain. Nowadays, researchers tend to investigate the computation in memory system based on single-array macro and mostly focus on the fully-connected structure. However, in practical applications, we must have multiple arrays or cores to run a more complicated neural network, such as the convolutional neural networks. The challenges would be different in multiple-array system with the single array case, and the convolutional operations are still inefficient in this novel computation in memory system.”​

In this research, a versatile memristor-based computing architecture was proposed for neural networks, and accordingly, eight 2K memristor crossbar arrays were integrated to implement the system. Especially, the team optimized the device stacks and developed a fabricating process which is compatible with current foundry process. The fabricated memristor arrays exhibit uniform multilevel resistive switching under identical programming conditions.

The photo of the hardware system with multiple memristor crossbar arrays. Credit: Huaqiang Wu’s Research Group

Mr. Peng Yao, the first author of this paper, comments: “When deploying a complete CNN into the multiple memristor arrays, the system performance would degrade due to the inherent device non-ideal characteristics within and between arrays. Conventional ex-situ training method could not address this problem with acceptable cost, and tuning all memristor weights by in-situ training method is hindered by the device nonlinearity and asymmetry and sophisticated peripheral modules.”​

The non-ideal device characteristics are considered as the substantial hurdles to result in the system performance degradation. To circumvent various non-ideal factors, a hybrid training method is proposed to implement the memristor-based CNN (mCNN). In the hybrid training, the ex-situ trained weights are firstly transferred to the memristor arrays, and in the next phase, only a part of the memristor weights are in-situ trained to recovery the system accuracy loss due to device non-ideal characteristics. In this paper, only the last FC layer is in-situ trained to reduce the hardware expense.

Meanwhile, in mCNN, the memristor-based convolutional operations are time-consuming due to the need to feed different patches of input during the sliding process. In this manner, the team proposed a spatial parallel technique by replicating the same kernels to different groups of memristor arrays. Different memristor arrays could deal with different input data in a parallel way and expedite convolutional sliding tremendously. The device non-ideal characteristics could incur the random transferring errors in different memristor groups regarding the same kernels, therefore, hybrid training method is adopted at the same time.

The system introduction. (a) Schematic of the system architecture with eight integrated memristor PE units and other functional blocks. (b) The structure of the five-layer mCNN for MNIST image recognition. Credit: Huaqiang Wu’s Research Group

The methods and experiments in the multiple memristor-array system are momentous for both fundamental studies and diverse applications. It suggests that for CIM system, the device non-ideal characteristics at the bottom level could be effectively addressed by the strategies at the system level. The proposed hybrid training method and spatial parallel technique at system-level have shown to be scalable to larger networks like ResNET, and they could be extrapolated to more general memristor-based CIM systems.

Professor Huaqiang Wu is positive and enthusiastic: “The hybrid-training method is a generic system-level solution that accommodates non-ideal device characteristics across different memristor crossbars for various neural networks, regardless of the type of memristive devices. Similarly, the spatial parallel technique could be generally extended to other computation in memory systems to efficiently enhance their overall performance. We expect that the proposed approach will enable the development of more powerful memristor-based neuromorphic systems, and finally revolutionize artificial-intelligence hardware”.
​

Novel memristor-enabled computation in memory architecture could revolutionize artificial-intelligence hardware | SciGlow

 

[Stranagor]

Shanghai Enlists AI in War Against Coronavirus

TANG SHIHUA
DATE : FEB 05 2020/SOURCE : YICAI

Shanghai Enlists AI in War Against Coronavirus

(Yicai Global) Feb. 5 -- Shanghai today officially launched its novel coronavirus pneumonia smart evaluation system that uses artificial intelligence and other technologies to help clinicians efficiently and accurately diagnose and make medical decisions about this deadly new disease.

Preliminary examination of more than 70 patients' conditions show that the results of this intelligent evaluation mechanism are highly consistent with those of experts at the city's Public Health Clinical Center, Shanghai Observer reported. The system can complete a quantitative analysis in only two to three seconds, whereas a manual evaluation by doctors can often take two to three hours, or even longer.

Local AI startup Yitu Technology developed this new medical tool, which applies the new technologies to analyze, evaluate and predict the prognosis of virus-induced lesions in lung CT images, under the guidance of the city's public health center, which commissioned it.

"Many employees are working into the wee hours every day," said Dr. Shi Lei, vice president of Yitu's healthcare division, referring to the more than 100 research and development staff who gave up their Spring Festival holiday and toiled day and night to bring the new breakthrough technology to fruition.

"This smart system will be popularized and used in designated hospitals across the country after more clinical verification," stated Shi Yuxin, deputy director of the public health center.

https://yicaiglobal.com/news/shanghai-enlists-ai-in-war-against-coronavirus