Artificial Intelligence thread
- Thread starter 9dashline
- Start date
The Tsinghua LLM group secured some funding.
Zhipu AI, one of China’s most promising challengers to OpenAI, has received funding from the country’s food delivery giant Meituan, which has a market cap of around $100 billion at the time of writing.
An affiliate of Zhipu AI recently added a Meituan subsidiary as its shareholder, which now owns a 10% stake in the firm, local media reported citing business filing information. The startup hasn’t disclosed its exact funding to date, only saying it raised “hundreds of million yuan” ($1 = 7.23 yuan) from a Series B round last September. Its investors include Qiming Venture Partners, Legend Capital and Tsinghua Holdings.
A multitude of Chinese companies are working to develop large language models (LLMs) that could potentially challenge their Western equivalents. One such company, Zhipu AI, hails from the academic realm, having spun out of the country’s prestigious Tsinghua University. Founded in 2019, the startup is led by Tang Jie, a professor in the university’s Department of Computer Science and Technology.
Zhipu recently open sourced its bilingual (Chinese and English) conversational AI model ChatGLM-6B, which is trained on six billion parameters and claims to be able to carry out inferences on a single consumer-grade graphics card, significantly lowering the cost of running an LLM. It also previously open sourced a more robust, general-purpose variant, the GLM-130B trained on 130 billion parameters. Its user-facing chatbot app ChatGLM is currently in a close beta phase, first targeted at academic and industry players.
Meituan’s investment came at a curious time. Just three weeks ago, the Chinese internet giant announced it would be acquiring Light Years Beyond, another prominent LLM player in China, for a hefty $234 million, despite the startup’s inception only four months prior. The change in ownership came after Light Years Beyond’s founder, Wang Huiwen, who’s also the billionaire co-founder of Meituan, announced his resignation from all corporate roles at the food delivery giant due to health reasons.
These investments are expected to give Meituan’s AI capabilities a big talent boost. In turn, the AI firms stand to gain by potentially tapping Meituan’s vast reach of 450 million users ordering food, buying groceries or booking hotels with the on-demand platform.
China’s OpenAI challenger Zhipu AI gets Meituan funding
Zhipu AI, one of China’s most promising challengers to OpenAI, has received funding from the country’s food delivery giant Meituan, which has a market cap of around $100 billion at the time of writing.
An affiliate of Zhipu AI recently added a Meituan subsidiary as its shareholder, which now owns a 10% stake in the firm, local media reported citing business filing information. The startup hasn’t disclosed its exact funding to date, only saying it raised “hundreds of million yuan” ($1 = 7.23 yuan) from a Series B round last September. Its investors include Qiming Venture Partners, Legend Capital and Tsinghua Holdings.
A multitude of Chinese companies are working to develop large language models (LLMs) that could potentially challenge their Western equivalents. One such company, Zhipu AI, hails from the academic realm, having spun out of the country’s prestigious Tsinghua University. Founded in 2019, the startup is led by Tang Jie, a professor in the university’s Department of Computer Science and Technology.
Zhipu recently open sourced its bilingual (Chinese and English) conversational AI model ChatGLM-6B, which is trained on six billion parameters and claims to be able to carry out inferences on a single consumer-grade graphics card, significantly lowering the cost of running an LLM. It also previously open sourced a more robust, general-purpose variant, the GLM-130B trained on 130 billion parameters. Its user-facing chatbot app ChatGLM is currently in a close beta phase, first targeted at academic and industry players.
Meituan’s investment came at a curious time. Just three weeks ago, the Chinese internet giant announced it would be acquiring Light Years Beyond, another prominent LLM player in China, for a hefty $234 million, despite the startup’s inception only four months prior. The change in ownership came after Light Years Beyond’s founder, Wang Huiwen, who’s also the billionaire co-founder of Meituan, announced his resignation from all corporate roles at the food delivery giant due to health reasons.
These investments are expected to give Meituan’s AI capabilities a big talent boost. In turn, the AI firms stand to gain by potentially tapping Meituan’s vast reach of 450 million users ordering food, buying groceries or booking hotels with the on-demand platform.
The Institute of Automation has developed a brain-like pulse neural network accelerator
Recently, the research group of Zeng Yi, a researcher at the Institute of Automation, Chinese Academy of Sciences, proposed an FPGA -based spiking neural network hardware accelerator "FireFly" , which integrates a DSP operation optimization strategy tailored to the characteristics of FPGA devices and an efficient synapse weight and membrane voltage memory access system that adapts to the spiking neural network data flow mode. It realizes the reasoning acceleration of spiking neural networks on hardware, and promotes the development of brain-like spiking neural networks towards practicality. Related research results were published in IEEE Transactions on Very Large Scale Integration (VLSI) Systems .
With the in-depth study of the algorithm mechanism, the performance of the brain-like spiking neural network has gradually improved, but the hardware adapted to the spiking neural network lags behind the development of the algorithm. As programmable hardware, FPGA is an ideal hardware carrier for emerging spiking neural networks. However, the existing spiking neural network accelerators for FPGAs are insufficient in both computing and storage efficiency.
"Zhimai Yinghuo " , as a high-throughput brain-like neural network accelerator with optimized computing and memory access, can effectively help solve the above problems. In order to improve the computing efficiency, this research uses the special computing module DSP48E2 in the Xilinx Ultrascale device to realize the efficient computing of the spiking neural network. In order to improve storage efficiency, the study designed a memory system to achieve efficient synaptic weight and membrane voltage memory access. At the same time, "Zhimai Yinghuo " is an accelerator for edge-type FPGA devices, and it can still achieve a peak computing throughput of 5.53TOP/s on FPGA devices with limited resources . In the existing research on SNN accelerators based on systolic arrays, "Zhimai · Yinghuo " has an 8.5 -fold increase in computing throughput compared with research using FPGA devices of the same magnitude (such as Cerebon [TVLSI'22] ) , enabling it to achieve millisecond-level delays in several deep spiking neural networks. As a lightweight accelerator, "FireFly " achieves higher computing efficiency than the existing spiking neural network accelerators using large FPGA devices.
"Zhimai · Yinghuo " is a phased achievement of the artificial intelligence platform " brain-like cognitive intelligence engine 'Zhimai'" (BrainCog) in the direction of software and hardware collaboration, laying a foundation for further layout of the application research of future spiking neural networks in more complex actual scenarios. In the next step, the team will continue to improve the performance of "Smart Firefly" in terms of hardware optimization, micro-architecture design, and sparse acceleration, and deploy it in real-world scenarios, such as autonomous visual positioning and navigation of smart cars based on pulse neural networks, high-speed obstacle avoidance of UAVs based on event cameras, and robot environment exploration in multi-task and multi - scenario situations .
The Cambrian Consortium won the bid for a 753 million yuan project and will supply and install intelligent computing hardware.
Cambrian issued an announcement stating that recently, the company (leader) formed a consortium with China Mobile Communications Group Zhejiang Co., Ltd. Taizhou Branch and Zhejiang Public Information Industry Co., Ltd. to participate in the bidding for the "Digital Infrastructure Improvement Project of Southeast Zhejiang Digital Economy Industrial Park (Phase I)". According to the "Announcement on the Results of the Digital Infrastructure Improvement Project (Phase I) of the Digital Economy Industrial Park in Southeast Zhejiang" issued by the Zhejiang Government Purchase Service Information Platform on July 25, 2023, the consortium formed by the company, China Mobile and Zhejiang Information is the winning bidder for the above-mentioned public bidding.
According to reports, Taizhou Huangyan Zhicheng Property Management Co., Ltd. is the purchaser of the Digital Infrastructure Improvement Project (Phase I) of the Digital Economy Industrial Park in Southeast Zhejiang. The consortium plans to win the bid with an amount of 753.081806 million yuan. According to the division of labor among the members of the consortium, the company, as the leader of the consortium, is responsible for the supply, installation and follow-up services of the intelligent computing hardware. It is estimated that the company's share of the total price is about 70% of the project.