The turmoil on the other side of the ocean is merely headline news in another country's newspapers.
It's so good that people forget it as soon as they see it.
Of course, the news of the first publicly revealed mage on Earth will be remembered for a long time.
However, Wiles hadn't had time to read the news in a long time, and he only noticed that the number in the upper left corner of the magic forum had been increased by one digit.
He realized that there was another monk in the world. After talking with his colleagues, he learned that the newly added monk was named Zheng Li, who was the richest man in China.
At the same time, he has an exceptional talent for scientific research.
These were just casual conversations during a meal.
After mastering magic, Wiles found that his appetite had increased significantly.
Research has found that their brains require more energy, thus needing more food to be digested in the stomach to support brain function.
After Wiles finished eating, he returned to his office with a premonition.
He will soon achieve a breakthrough in the field of artificial intelligence.
This method is absolutely feasible.
Previously, scientists proposed a novel neural model for achieving associative memory, which is based on a hierarchical generative network that receives external stimuli through sensory neurons.
This novel neural model is trained using predictive encoding.
Predictive coding is an error-based learning algorithm inspired by information processing in the human cerebral cortex.
The associative memory model draws inspiration from the way neurons in the human brain store, associate, and retrieve information.
Due to its importance in human intelligence, computational models of associative memory have existed for decades.
It includes an autoassociative memory that allows storing and retrieving data points when noise or partial variations are provided, and a heterogeneous associative memory that can store and recall multimodal data.
However, this remained only at the conceptual stage; at the time, it was merely a proposal.
It is speculated that the desired purpose can be achieved through a new type of neural network.
However, it has not yet been implemented in practice.
Wiles, after conducting artificial intelligence learning, believes that this approach has the greatest potential.
Of course, there are also routes with greater potential, but those routes with greater potential are constrained by hardware limitations.
Hardware breakthroughs are much more difficult than software breakthroughs.
For example, there are learning automata. This type of automaton can learn from data by changing the probability of its own state transition function based on the learning data, and can use its own state to encode information.
Unlike neural networks, learnable automata naturally possess the ability to encode data in a time sequence and have good interpretability.
However, the problem with this approach is that the algorithm path is too complex, with too many states, making it difficult to implement in hardware.
Upload or not?
"yes'
After a long period of research, development, and testing, they finally created a relatively complete artificial intelligence.
It has also passed the internal Turing test.
Wiles felt a pang of anxiety after clicking the upload button.
Continue read on readnovelmtl.com
