In 2017, artificial intelligence provided a driving force to accelerate the rhythm of the entire industry, and products emerged in an endless stream. Various manufacturers continued to innovate and bring more vitality to the entire market.
In 2018, it was reputed as the first year of artificial intelligence. In the coming years, it demonstrated its degree of smart promotion. If you have the electrical appliances used to observe everyday life, you can see that many products have changed their appearance and have a new look. appearance.
TVs in the home have more products with 8K, equipped with smarter speakers, synchronized audio and video, stunning you, home fridges, video panel design, and more entertaining designs. Get up; Even the mobile phones that we rely on most of the day-to-day market, the thousand-thousand-dollar and full-screen series in the whole market, make you really feel that this era is changing all the time.
MIT Designs Low-Power Artificial Intelligence Chips
As smart phones become more and more intelligent, there are more and more artificial intelligence services such as digital assistants and real-time translation. The neural networks for computing these services are usually in the cloud, and the data on mobile phones are also transmitted back and forth in the cloud. Therefore, its operation requires a large amount of communication bandwidth as a support.
However, as a neural network of a graphics processor, a large amount of energy is required in normal operation, and the limited battery power becomes a major factor for hindrance.
To this end, MIT engineers designed a chip that could significantly reduce the need to transfer data back and forth between the chip's memory and the processor, thereby reducing power consumption by 95%. The new chip is another way to use analog circuits, Calculating all input in parallel in memory greatly reduces the amount of data that needs to be pushed, saving a lot of energy.
Although big data does make our life much easier, in fact, in the data operation, each input is a process that requires a lot of data to move, so it will consume a lot of energy to maintain normal operation. Although previous researchers were Reducing the power consumption of the neural network and creating a chip for processing data in memory, but this is the new chip, this is the first time that this method is used to run convolutional neural networks based on image-based artificial intelligence applications. It is also very wide.
Artificial intelligence chips will have great use in different fields
The introduction of artificial intelligence has made ordinary families feel the rapid development of science and technology. How to make smart phones, household appliances, and various Internet of things devices equipped with artificial intelligence is also driving the industry's big names to switch to low-power labor. Smart chip.
Apple, which has been at the top of the mobile phone industry, has integrated its Neural Engine chip into the iPhone X, including enhancements to its facial recognition technology. ARM has earlier released a new chip, the ARM machine. Learning processor, the chip is mainly targeted at artificial intelligence tasks, from translation to facial recognition.
The other chip, Qualcomm, has also introduced the latest mobile chip, the Snapdragon 845, which is equipped with a graphics processor and artificial intelligence as a top priority. Not only that, it also released the Snapdragon 820E chip. The main targets are drones, robots and industrial equipment.
In addition to chip companies introducing more intelligent products, IBM and Intel are also laying out artificial intelligence chips. They will develop a neuromorphic chip whose architecture is inspired by the human brain and its amazing energy efficiency. , you can run a powerful machine learning, the energy consumption is only a very small part, can be described as very energy-efficient.
In the future life, products will be more intelligent and convenient. Cloud computing will be combined with big data. The era of the Internet of Everything will gradually emerge, and the family's entertainment life will become more sci-fi and richer. At the same time, it will also solve the problem of energy shortage. progress.
Although the current situation is that the real artificial intelligence has a long way to go. However, with the constant speed of innovation, I believe we will not wait too long!
