Sep 10 2020
MobiusTrend, a market research organization in Hong Kong, recently released a research report on 'The Core Algorithm of Google and WIMI Helps the Development of Hologram AI Vision'.
Google has proved that quantum computing can also solve the problem of image classification in traditional machine learning, and with the development of technology, quantum computers will surpass classical neural networks in learning capabilities. In addition, quantum computing can also solve some thorny problems in classical networks, such as preventing gradient disappearance in model training.
The researchers said that this is to show that quantum mechanics textbooks can reveal artificial intelligence problems, and believe that the maximum achievable classification accuracy is if an algorithm must make a decision after discovering the first "quantum" light (i.e., photons) passing through an LCD screen to display an image from a data set. MNIST, the most classic calculation that can be done is to detect a photon falling on an image pixel, and then to guess the number from the distribution of light intensity by resetting the brightness of each image to one unit sum.
Holographic technology may be the key to quantum computing
Breakthroughs in light research may be the ticket to the future of quantum computing. Using the quantum nature of the interaction between electrons and light, beams are separated by energy rather than space, which allows them to use light pulses to encrypt information about electronic wave and map it with high-speed electron microscopes.
Photogrammetry measures the degree to which different colors of light are illuminated to a photographic film. Light, however, is also a wave and is characterized by phases. The phase specifies the position of the points within the wave period and is related to the information depth, which means that the phase of the light scattered by an object can be recorded to retrieve its complete 3D shape, which cannot be obtained from a simple photograph. This is the basis of optical holography, which is displayed by fancy holograms in science fiction movies such as Star Wars.
WIMI provides the one-stop service of holographic AR technology. Its business covers from holographic visual AI synthesis and presentation, holographic interactive software development, holographic AR advertising, holographic AR SDK pay, 5G holographic communication software development to holographic face recognition. Meanwhile, its business application scenarios are mainly gathered in five professional fields: home entertainment, light field cinema, performing system, business distribution system, and the advertising display system.
At present, traditional holography can extract 3D information by measuring the difference in the distance traveled by light from different parts of the objects, but this requires additional reference beams from different directions to measure the interference between the two. The concept is the same as that of an electron, but because the wavelength is much shorter, we can now achieve a higher spatial resolution. For example, we can record a holographic film of a fast-moving object by using ultrashort electronic pulses to form a hologram.
In addition to quantum computing, the technology has the highest spatial resolution compared to alternatives and has the potential to change the way we think about light in our daily lives. So far, science and technology have been limited to freely propagating photons for use in macroscopic optical devices, and the new holographic technology allows us to see what happens to nanoscale light. This is the first step toward miniaturizing and integrating optical devices into integrated circuits.
Communications industry experts pointed out that radio frequency is particularly critical in the design of 5G mobile phones. The highest manufacturing costs of 4G mobile phones are on the screen and processor, but the highest costs of 5G mobile phones may shift to a complete radio frequency package. Navian predicts that the market size of RF front-end chips in mobile terminals alone will reach $21.2 billion in 2020, with a compound annual growth rate of 15.4 percent. In the 5G era, more frequency band resources will be put into use. Multi-mode and multi-frequency will increase the demand for RF front-end chips. Meanwhile, key technologies such as Massive MIMO, beamforming, carrier aggregation, and millimeter-wave will also increase the demand for RF front-end chips and directly promote the growth of the RF front-end chip market.