Netease Technology News on April 7th, Max Versace, CEO of Boston deep learning company Neurola, wrote an article on VentureBeat, a technology blogging website, saying that in 2017, artificial intelligence (AI) oximes îˆƒé” îˆƒé” îˆƒé” îˆƒé” ç«½ç ¸ II 裉斓 裉斓 裉斓 裉斓 裉斓 裉斓 娜 娜 娜 II II II
The full text of the article is as follows:
Let us pause for a second and look at the birds, squirrels and even insects outside the window. These organisms are performing complex tasks, including sensing food and threats, patrolling around trees, and tracking or evading other animals. These worms and small animals can easily do things. At present, there is no robot or drone on earth that can do it.
Although "natural" intelligence is rich and varied, today's artificial intelligence is still very primitive. At present, artificial intelligence tools are still "semi-finished products." Their design and construction (programming) is mainly aimed at a special target and is relatively simple. For example, codes and sensors that manage drones, driverless cars, and toys often focus on just one of many tasks that need to be performed—navigation, target recognition, or speech recognition.
Because these applications are similar to biological organisms, artificial intelligence requires a "brain." The current design and construction of artificial intelligence aims to achieve a narrow, independent function. You can call this "flue" function, where each function is disconnected and independent during processing. This means that artificial intelligence can defeat the human chess champion, but it tends to collapse when faced with a new scene. Unlike the artificial intelligence chess players, human chess champions can not only finish the game with high quality, but also perform other activities that require a lot of processing and judgment, such as standing up, driving, chatting with his children. , listening to music, drawing, etc. So, when you directly compare the capabilities of artificial intelligence with the brain functions of humans and even animals, you can see that we have a long way to go in artificial intelligence. In the era of science and technology, we are still in the stage of "Caveman".
The main difference between the artificial intelligence we have today and the artificial intelligence of 2027 is that the latter will behave more like the brains of humans and animals, and its capabilities will far exceed today's artificial intelligence. The process of processing life information on the earth is not "fluo" but uses a variety of senses; we humans can make decisions based on multiple complex factors at the same time to help us achieve the best results.
Take the mouse's brain as an example. Even the smallest animal, its brain has evolved enough to solve complex problems such as finding food, avoiding predators, and interacting with other animals. The mouse's brain is only about two grams, but it can integrate navigation functions with sight, smell, and touch (through beard) functions, allowing tasks such as sensing, planning, navigation, and avoiding obstacles to be accomplished. In the mouse's brain, these separate functions can eventually be integrated to provide a "total solution" for the task faced. The secret of animals (and humans) brains is that they find a way to run these skills together in the same low-power containers.
This "co-operation" is what we call the "whole brain" approach, and this new model is the direction of artificial intelligence. "Integrated processing" will become commonplace, and the boundaries between software, artificial intelligence, and human (animal) wisdom will become blurred. Just as human or animal brains operate effectively and autonomously, relying on and integrating multiple areas of the brain, tomorrow's artificial intelligence will use integrated, in-depth learning methods and edge processing functions to increasingly solve problems in real time.
By embedding multiple artificial intelligence in the same container or in a single computing module, the artificial intelligence system will achieve better and faster performance due to the synergy between the systems. This will allow artificial intelligence to complete abstract reasoning processes, allowing machines to perform complex, non-intuitive operations that will bring them closer to ours.
For example, the line between artificial intelligence and software is becoming increasingly blurred, making transportation easier and safer. Today's driverless cars use a "flue" approach, adding one sensor or module at a time, and then integrating all of these processing information streams in hopes that they will work. And humans can co-integrate multiple tactical plans - not just "Look! There is a pit in front!" - This includes:
(1) High-level navigation: "I know there are buildings there; è¦å½Ç” 谡饫 谡饫 谡饫  !
(2) Avoid collisions remotely: "The car is dangerous. I will keep the distance."
(3) High-level planning: "I'd better go that route because it may have fewer vehicles."
Although the traditional approach will lead to uncontrollable conflicts between hard-to-integrate software and hardware components, the "whole-brain" artificial intelligence approach uses artificial neurons that possess simulated synapses that process information in much the same way as the human brain.
Another example is our cooperation with NASA. When Neurala collaborated with NASA to design a “rat brain†and guided the Mars Rover in a simulated Mars environment, we followed this “whole brain†approach because we could only rely on a small amount of The computing power can only provide one solution - it not only needs to integrate all these functions in a single container, but also it can be implemented effectively.
Even today's "pipelined" artificial intelligence is making software and machines that communicate with us better, and has increased productivity in many aspects of our lives. As AI begins to imitate advanced human and animal brain activity, it will become an increasingly useful tool that can solve problems in real time and has human decision-making capabilities. In the next 10 years, this kind of integrated treatment that can make the insignificant mouse look like a genius will be an artificial intelligence that can benefit all people.
(Liu Chun)
Product Name: Car Charger
Place of Origin: Guangdong, China (Mainland)
Brand Name: OEM
Output Type: DC
Connection: Other
Rated Voltage: 12V-24V
Working Temp: 0-55℃
Weight: 36g
Materials: PC+ABS
Color: White Black
Warranty: 1 year
Suitable for:Most digital devices
SMART PROTECTION & ATTRACTIVE DESIGN ------ Intelligent circuit design protects against short circuiting,over-heating,over-current,and over-charging. Charging stops when battery is full. Car charger with blue LED indicator,which makes it convenient to find exactly where the connection should go; And the light is soft enough not to distract at night.
Dual USB Car Charger Adapter,USB Smart Port Charger ,Car Charger,USB Car Charger For Phone
Shenzhen Waweis Technology Co., Ltd. , https://www.huaweishiadapter.com