Huawei. Pangu big model, the real uphill of Chinese chips.
Huawei released Pangu Model 3.0 to provide intelligent solutions for industries. Zhang Pingan, Executive Director of Huawei and CEO of HUAWEI CLOUD, released Pangu Model 3.0 at Huawei Developer Conference 2023. This model is Huawei's intelligent solution for the industry, aiming to provide customized AI services for all walks of life.
Pangu Grand Model 3.0 adopts a three-tier architecture of 5+N+X and provides a series of large models to meet the diverse skill needs of different industries. The L-zero Pangu model provides basic models such as natural language understanding, visual processing, multimodal processing, predictive analysis, and scientific computing. Customers can choose models of different sizes to meet the needs of different scenarios, different responses, speeds, and different delays.
The Pangu large model on the first floor of L provides general models in various industries, including large models in the fields of government affairs, finance, manufacturing, mining, meteorology, etc., and customers can conduct secondary training based on these general models and customize their own exclusive models. In the L2 layer, Pangu Large Model further provides more detailed scene models, providing customers with out-of-the-box model services for specific business scenarios.
The design of the Pangu model is completely decoupled, so that the model can quickly adapt to the changing needs of the industry. Customers can load independent datasets for their own large models, or upgrade the base model or capability set separately. HUAWEI CLOUD also provides a large model industry development kit, on which customers can have their own exclusive industry large models through secondary training on their own data.
In addition, the deployment form of the Pangu model is also very flexible, supporting multiple deployment methods such as public cloud, private cloud, and hybrid cloud to meet customers' data security and compliance requirements. The release of the Pangu Grand Model has brought great value and change to all walks of life. In the field of government affairs, through Xiaofu, a smart assistant based on the Pangu government affairs model, the government can better understand the consultation intentions of the people, provide personalized government services, and improve service efficiency and satisfaction.
In the field of coal mining, the application of the Pangu Mine Grand Model enables coal miners to work on the ground, improving the safety and comfort of the working environment.
In the railway field, the Pangu Railway Big Model can quickly and accurately identify truck faults, provide reliable digital assistants for freight train inspectors, and improve freight efficiency and safety. In the field of meteorology, the Pangea Meteorological Big Model provides more accurate and rapid weather prediction results through Al inference, providing better support for meteorological research and services.
In the financial field, the application of Pangu Financial Model enables counter staff to handle customer needs more efficiently and improve service quality and efficiency.
In the field of manufacturing, the Pangu manufacturing model can quickly generate production plans, improving production efficiency and accuracy. In the field of drug discovery, the application of Pangu drug molecule model shortens the drug development cycle and cost, and provides more efficient support for new drug research and development.
The release of the Pangu model marks Huawei's further development in the field of artificial intelligence, bringing more intelligent and efficient solutions to all walks of life, and promoting the digital transformation and innovation of the industry.
Scientists using ferroelectric polymers to make artificial muscles for the first time at Pennsylvania State University-led international research team have discovered a new type of ferroelectric polymer capable of converting electrical energy into mechanical strain and has the potential to become a high-performance mechanical strain controller or brake.
The research team solved the limitations of traditional piezoelectric materials in terms of mechanical strain by developing an osmotic ferropolymer nanocomposite. They introduced nanoparticles into the ferroelectric polymer and formed a network of interconnected poles inside the material. This design allows the ferroelectric phase transition to be induced at a lower electric field, requiring less than 10% of the electric field strength required by traditional ferroelectric materials.
By using an electrothermal method heated by Joule, the researchers were able to drive the phase change of the nanocomposite polymer. This method not only increases the force of the soft material, but also reduces the requirements of the drive field, making this new ferroelectric polymer manoeuvrable in a wider range of applications.
The new ferropolymers are highly flexible and environmentally adaptable, and have great application potential in medical devices, advanced robotics and precision positioning systems. By applying new ferropolymers to soft robots, higher strain capacity and mechanical energy density can be achieved. This material can be closer to the characteristics of human muscles, mimicking human muscles, and has advantages in motion control and motion simulation.
For soft robot researchers, the development of new ferroelectric polymers has brought new possibilities for the design of soft robots. In addition, the new ferropolymers have the advantages of lower cost and lower weight, which is of great significance for the growing field of soft robotics.
By developing this new material, the researchers provide solutions to challenges in the field of soft material drive, promoting the development and application of soft robotics. Portable solar water collectors can solve the problem of water shortage.
Researchers at the University of California, Berkeley, have designed an extreme weather-proven handheld device capable of extracting water molecules from the air and converting them into drinking water using only ambient sunlight as an energy source. The researchers used metal-organic frameworks, ultra-porous materials, which are known as ultra-porous materials with a high degree of adjustability and large internal surface area, as a key technique.
The study used a specially designed MOF for multi-organ failure and accelerated and scaled up the design of these molecules, materials, and devices through data science and machine learning methods. By utilizing the special structure of MOF, the device can efficiently extract water molecules from the air and convert them into drinking water.
In addition, the research team has achieved sustainable, emission-free access to water by using solar energy as an energy source to power the device. The results of this study show that this handheld device has efficient water molecule extraction capabilities, which can convert 85 to 90 percent of atmospheric water vapor into drinking water. Tests in some of North America's hottest and driest regions, such as California's Death Valley National Park, have shown that the device can effectively deliver clean drinking water under extreme conditions.
