分类： 计算机科学 >> 计算机科学技术其他学科 提交时间： 2023-12-18
摘要： In this paper, we propose a new approach to building a artificial general intelligence with selfawareness, which includes: (1) a new method to implement attention mechanisms; (2) a way to givemachines self-demands; (3) how to form a value evaluation system compatible with the network; (4) away to create the world models; (5) how to realize a top-down, hierarchical thinking decision-makingchain; (6) a way to achieve general decision-making and response capabilities; (7) a way for a machineto directly obtain human experience through language. In the paper, we first analyze some of the shortcomings of current LLMs (Large Language Model) and propose ideas for improvement. Then weanalyze why our scheme can solve the above problems and provide detailed steps for implementingour scheme. In chapter 6, we analyze the advantages and disadvantages of our scheme and proposefurther research directions. Finally, we propose our thoughts on the next step of AI development.
分类： 计算机科学 >> 计算机科学技术其他学科 提交时间： 2023-05-06
摘要： With a preliminary exploration of the capability boundaries of LLM(Language Large Model),we believe that the current mainstream artificial intelligence generally adopts the technical of attention mechanism + deep learning + reinforcement learning, which cannot be applied to those fields that are difficult to a lot of trial and error. So, to achieve AGI (Artificial General Intelligence) that works in any field, its better to change the way we do it. Therefore, we propose a set of machine learning solution different from deep learning + reinforcement learning. It adopts small samples and cumulative learning, and also realizes the attention mechanism similar to transformer, and also creates a fully connected knowledge network. In addition, it can realize interactive decision making with the environment without using lots of trial and error style learning. In addition, humans can preset different innate needs to it to achieve multi-objective balance, thus achieving far higher security than the current artificial intelligence. In this paper, we propose a set of new machine learning techniques which maybe guide humans realizes AGI
分类： 光学 >> 量子光学 提交时间： 2023-02-19
Qi Shen
Jian-Yu Guan
Ji-Gang Ren
Ting Zeng
Lei Hou
Min Li
Yuan Cao
Jin-Jian Han
Meng-Zhe Lian
Yan-Wei Chen
Xin-Xin Peng
Shao-Mao Wang
Dan-Yang Zhu
Xi-Ping Shi
Zheng-Guo Wang
Ye Li
Wei-Yue Liu
Ge-Sheng Pan
Yong Wang
Zhao-Hui Li
摘要： Optical clock networks play important roles in various fields, such as precise navigation, redefinition of "second" unit, and gravitational tests. To establish a global-scale optical clock network, it is essential to disseminate time and frequency with a stability of $10^{-19}$ over a long-distance free-space link. However, such attempts were limited to dozens of kilometers in mirror-folded configuration. Here, we take a crucial step toward future satellite-based time-frequency disseminations. By developing the key technologies, including high-power frequency combs, high-stability and high-efficiency optical transceiver systems, and efficient linear optical sampling, we demonstrate free-space time-frequency dissemination over two independent links with femtosecond time deviation, $3\times10^{-19}$ at 10,000 s residual instability and $1.6\times10^{-20}\pm 4.3\times10^{-19}$ offset. This level of the stability retains for an increased channel loss up to 89 dB. Our work can not only be directly used in ground-based application, but also firmly laid the groundwork for future satellite time-frequency dissemination.
点击量
待评议
点击量
下载量
评论
