[Guide]: Google's acquisition of deepmind can be described as cost-effective, AlphaGo is a masterpiece of deepmind, but this is a past tense. In the early morning, deepmind announced a new breakthrough, training the AI ​​of the human brain model, and learning to cut the way like a mammal. Google's deepmind UK team published a paper in the world's top academic journal Nature, which recently developed an artificial intelligence program with mammal-like pathfinding capabilities, much like the workings of grid cells in the brain. Starting from home to a new location, returning to the original route, and choosing the best possible shortcuts, this is a simple task that most animals can do. However, the brain's instinctive navigation mechanism is not fully understood. Scientists have found three cells associated with pathways in animals and human brains: positional cells, directional cells, and grid cells. Position cells can discharge when the subject reaches a specific location, giving memory to past locations; directional cells can sense the direction of advancement; grid cells are the most mysterious one: they can divide the entire space into honeycombs. A grid of edges, like the coordinate system on a map. The Mosul couple who discovered the grid cells won the 2014 Nobel Prize in Physiology or Medicine. However, is grid cell just providing GPS location services in a space environment? Some scientists speculate that they will also participate in vector calculations to aid animal planning pathways. The DeepMind team decided to use artificial neural networks to test the above conjecture. Artificial neural network is an operational structure that uses multi-layer processing to simulate a neural network of the brain. The team first used the deep learning algorithm to train the neural network to learn the mammalian foraging motion path, and used the linear velocity, angular velocity and other signals to locate in the visual environment. The researchers then discovered that a structure similar to the characteristics of grid cell activity was born! In previous training, the researchers did not deliberately guide the neural network to produce such a structure. A grid structure similar to rat grid cells (bottom) appears automatically in artificial neural networks (top) This again shows the gratifying and awesome aspect of deep learning: this is an algorithm that trains the machine to self-explore through a large number of matching input and output values, and the resulting machine logic is a black box that is unknown. Just as AlphaGo automatically comprehends the unprecedented "flying fairy" in the history of the human millennium, this neural network also automatically presents a surprising structure. The DeepMind team then used reinforcement learning to test whether the grid structure could be vector navigated. Reinforcement learning is commonly used to train game AI. Humans tell AI a game's scoring and reward mechanism, but they don't teach game methods. AI evolves itself in the process of repeatedly playing games and striving for higher scores. Later AlphaGo completely abandoned the human game experience and evolved from zero to a stronger version in a pure self-game. The researchers combined the previously emerging grid structure with a larger neural network architecture into an artificial intelligence, placed in a virtual reality game environment. After undergoing intensive learning, the artificial intelligence's ability to navigate to the destination in the game maze is beyond the average person and reaches the level of the professional game player. It can look for new routes and shortcuts like mammals. Artificial intelligence learning to cut corners in the maze The most important thing is that when the researchers "silently" the original grid structure, the navigation ability of the artificial intelligence body will be weaker, and the distance and direction of the target are more inaccurate. Dharshan Kumaran, one of the authors of the paper, said: "We have demonstrated that grid cells are much more than just providing GPS positioning signals, but also a core navigation mechanism that the brain relies on to calculate the shortest distance between two locations." Unlike Yann Lecun, the chief scientist of Facebook's artificial intelligence, who has always emphasized "artificial intelligence is not bionics," DeepMind, the genius founder of Demis Hassabis, is keen to explore the complementarity between artificial intelligence and brain science. This research once again embodies their scientific philosophy: artificial intelligence algorithms inspired by brain science can in turn help humans explore the brain's operating mechanism and thus better understand the inherent logic of artificial intelligence. Hassabis commented: "We believe that artificial intelligence and neuroscience are mutually inspiring. This work is a good proof: by developing an artificial intelligence body that can navigate in complex environments, we are There is a deeper understanding of the importance of mammalian navigation." The DeepMind team believes that similar research methods can also be used to explore the mechanisms of brain hearing and control of the limbs. In the far future, neuroscientists can even experiment with artificial intelligence instead of mice.
ZGAR AZ Vape Pods 5.0
ZGAR electronic cigarette uses high-tech R&D, food grade disposable pod device and high-quality raw material. All package designs are Original IP. Our designer team is from Hong Kong. We have very high requirements for product quality, flavors taste and packaging design. The E-liquid is imported, materials are food grade, and assembly plant is medical-grade dust-free workshops.
From production to packaging, the whole system of tracking, efficient and orderly process, achieving daily efficient output. WEIKA pays attention to the details of each process control. The first class dust-free production workshop has passed the GMP food and drug production standard certification, ensuring quality and safety. We choose the products with a traceability system, which can not only effectively track and trace all kinds of data, but also ensure good product quality.
We offer best price, high quality Pods, Pods Touch Screen, Empty Pod System, Pod Vape, Disposable Pod device, E-cigar, Vape Pods to all over the world.
Much Better Vaping Experience!
ZGAR AZ Vape 5.0 Pods,ZGAR AZ Vape Pods 5.0,ZGAR AZ Vape Pods 5.0 Pod System Vape,ZGAR AZ Vape Pods 5.0 Disposable Pod Vape Systems ZGAR INTERNATIONAL(HK)CO., LIMITED , https://www.oemvape-pen.com
