西安夜生活论坛

Why is the game fun? Author: Wang Yahui, Version: Turing New Knowledge | People’s Posts and Telecommunications Publishing House, March 2023. (Poster design: Liu Xiaofei)

Recommended reason:

On May 12th this year, Nintendo released "the legend of zelda: Tears of the Kingdom", which sold 10 million copies worldwide in just three days, and triggered a wave of sharing among players on social media at home and abroad. With the development of media technology, the "ninth art" of games is increasingly attractive to people. Why is the game so fun? As a veteran gamer and the author of "The Game of China", Wang Yahui answers this fundamental question in a simple way through a large number of classic game cases such as Super Mario, Tetris, Diablo and the legend of zelda.

This book holds that games have clear and macroscopic "Rules" written in the instructions, and also have hidden and precise "Game Mechanics" that require players to explore themselves, while classic games can often derive the most complex and changeable mechanisms with the simplest rules. Judging from the two dimensions of time and space, which are the most basic for human beings to know the world, whether it is an extremely simple but popular elimination game such as Tetris and Bubble Dragon, or the most popular Jedi Survival, it is a superb use of the "space restriction" mechanism. Max payne, a classic adventure action game, tapped the potential of time mechanism through the use of bullet time.

In addition to space and time, this book also discusses the game mechanism in many aspects, such as money, props, skills, tasks, collection, etc., sorts out the classic design patterns in the study of game mechanism, and expands the horizons of ordinary readers to understand the game.

Editor/Shang Zhongming Wang Mingbo

Proofreading/Xue Jingning Fu Chunyu

CVPR 2023, one of the top three computer vision conferences, was exposed in the paper of China Smart Car Company.

Domestic autonomous driving companyHao mo zhi hangThe paper Bev-Lane Det: A Simple and Effective 3D Lane Detection Baseline submitted by the technical team of its Artificial Intelligence Center was successfully selected into CVPR2023.

What problems should the new research solve?

3D lane detectionAs a hot topic in the field of automatic driving, it plays a vital role in vehicle routing. In the past, related technologies were often difficult to be applied because of its complex spatial transformation and inflexible representation of 3D lanes.

Faced with these problems, Millie independently developed an efficient and robust monocular 3D lane detection, referred to as:BEV-LaneDet.

In BEV-LaneDet, Millimeter team first converted all the internal and external parameters of the input image into unified internal and external parameters through the virtual camera, which ensured the consistency of the spatial relationship of the front cameras in different vehicles and effectively promoted the learning process.

Moreover, Millimeter Team uses feature extractor to extract the features of the front view image, and proposes a simple but effective 3D lane representation, which is called "Key-Points Representation". This module is more suitable for representing complex and diverse 3D lane structures.

On this basis, inspired by FPN, Millicent team also designed a lightweight and chip-friendly spatial transformation module called "Spatial Transformation Pyramid", which is a fast multi-scale spatial transformation module based on VPN, mainly responsible for the transformation from front view features to BEV features.

Finally, the team predicted the lane on the plane tangent to the local road surface, divided the local road surface into s1×s2 grids, predicted confidence, embedding for clustering, lateral error offset and the height of each grid, and used fast clustering method to fuse the results of each branch in reasoning to obtain 3D lanes.

How did BEV-LaneDet perform?

Through the above methods, the F1-S?core of BEV-LaneDet can reach 185FPS on the OpenLane dataset, which is 10.6% higher than the current SOTA.

BEV-LaneDet can also lead the current SOTA method by 4.0% in the performance of Apollo 3D Lane Synthetic.

At present, Mimo Zhixing is the first echelon player in domestic self-driving mass production, and the paper written by its technical team BEV-LaneDet was selected for CVPR, which is actually the embodiment of the technical strength behind the commercial progress.

At present, HPilot has iterated three generations of products, and has achieved mass production on nearly 20 models of Wei brand, tank, Haval, Euler, Great Wall Gun and other brands. In the future, passenger cars equipped with Millimeter-assisted driving products will reach millions of orders of magnitude. As of February 2023, the mileage of user-assisted driving has exceeded 35 million kilometers. NOH, the first city in China that can land on a large scale, has also reached the delivery status with its software sealed.

In addition, the terminal logistics automatic delivery vehicle has initially completed the commercial closed loop, delivering more than 1,000 units, and the delivery order volume of the terminal logistics automatic delivery vehicle Little Magic Camel series has also exceeded 160,000, and the commercialization process is accelerating in an all-round way.

Introduction to CVPR

IEEE (Institute of Electrical and Electronics Engineers) CVPR (Conference on Computer Vision and Pattern Recognition) is one of the three top conferences in the field of computer vision and pattern recognition.

The main content of the conference is computer vision and pattern recognition technology. CVPR is the world’s top computer vision conference (one of the three top conferences, the other two are ICCV and ECCV). In recent years, there are about 1,500 participants every year, and the number of papers collected is generally about 300.

CVPR has strict employment standards. According to CVPR officials, CVPR2023 received a total of 9,155 papers, a record high, and the organizing committee finally hired 2,360 papers, with an acceptance rate of 25.78%.

CVPR is also the witness of major changes in the field of AI vision. The current wave of AI driven by deep learning originated from Hinton’s blockbuster in CVPR Challenge.

Address: https://arxiv.org/pdf/2210.06006.pdf.

— finish —

[Smart Car Reference] Original content is prohibited from being reproduced at will without the authorization of the account.

Inspired by recent biological discoveries, researchers have developed a new model that shows enhanced memory performance. This is achieved by modifying a classic neural network. Computer models play a key role in studying the process of making and retaining memories and other complex information in the brain. However, building this model is a delicate task.

The intricate interaction between electrical and biochemical signals and the connection network between neurons and other cell types create the basic structure for the formation of memory. Nevertheless, due to the limited understanding of the basic biology of the brain, it has proved to be a difficult task to code the complex biology of the brain into a computer model for further study.

Researchers at Okinawa Institute of Science and Technology (OIST) improved the widely used memory computer model (called Hopfield Network) by incorporating biological insights. This change has inspired a neural network, which not only better reflects the connection between neurons and other cells in the brain, but also has the ability to store more memories.

Thomas Burns, a doctoral student in the group of Professor Zhishu Shenjing, who is the head of OIST’s neurocoding and brain computing department, said that the increased complexity in the network makes it more realistic.

"Why is there so much complexity in biology? Memory may be one reason, "Mr. Burns said.

In the classical Hopfield network (left), each neuron (I, J, K, L) is connected with other neurons in pairs. In the improved network made by Burns and Professor Shenjing, three or more groups of neurons can be connected at the same time. Source: Thomas Burns (OIST)

Hopfield network stores memory as a weighted connection pattern between different neurons in the system. The network is "trained" to encode these patterns, and then researchers can test its memory of these patterns by presenting a series of vague or incomplete patterns to see if the network can recognize them as patterns it already knows. However, in the classical Hopfield network, the neurons in the model are connected with other neurons in the network to form a series of so-called "paired" connections.

Paired connections represent the connection between two neurons at the synapse, which is the connection point between two neurons in the brain. But in reality, neurons have complex branching structures called dendrites, which provide multiple connection points, so the brain relies on more complex synaptic arrangements to complete its cognitive work. In addition, the connections between neurons are regulated by other cell types called astrocytes.

Burns explained: "There are only paired connections between neurons in the brain, which is simply unrealistic. He created an improved Hopfield network, in which not only pairs of neurons, but also three, four or more groups of neurons can be connected, such as astrocytes and dendritic trees in the brain. "

Although the new network allows these so-called "collective" connections, on the whole, it contains the same number of connections as before. The researchers found that a hybrid network with paired connections and collective connections performed best and retained the most memory. They estimate that its effect is more than twice that of the traditional Hopfield network.

It turns out that you actually need to balance the combination of various features to some extent, Burns said. A single synapse is necessary, but you should also need some dendritic trees and some astrocytes.

Hopfield networks are very important for simulating brain processes, but they also have powerful other uses. For example, a very similar network type called Transformers is a language tool based on artificial intelligence, such as ChatGPT, so the improvements identified by Burns and Professor Shenjing may also make such tools more powerful.

Burns and his colleagues plan to continue to study their modified Hopfield networks to make them more powerful. For example, in the brain, the connection strength between neurons is usually different in two directions, so researchers want to know whether this asymmetric feature can also improve the performance of the network. In addition, he also wants to explore ways to make the memories of the network interact, just as they do in the human brain. Our memory is multifaceted and huge. We still have a lot to discover. "

Live on March 12 th, in the 27th round of Premier League, Tottenham beat Nottingham Forest 3-1 at home. With this victory, Tottenham consolidated its position in the top four of the standings. After the game, Sun Xingyi, who scored a goal, expressed his joy of winning through social media.

Sun Xingyi wrote: "It’s a beautiful day to get three goals and three points. Have a good rest and recover, and fight again next week. "

At 11pm on Saturday, March 18th, Beijing time, in the 28th round of Premier League, Tottenham Hotspur will challenge Southampton away.

The definition of AI artificial intelligence

In recent years, artificial intelligence can be said to be really hot, which has attracted the attention of many people and enterprises. It is a branch of computer science, which can imitate human intelligence to perform tasks, which is equivalent to having self-thinking consciousness. It is a new technical science used to research, develop, simulate, extend and expand the theoretical technology and application system of human intelligence. Since its appearance, artificial intelligence can be said to be increasingly mature and its application fields are increasing. Artificial intelligence can simulate the information process of human consciousness and thinking, and can deal with problems like human beings, even surpassing human beings.

Application of AI artificial intelligence in enterprises

Artificial intelligence can now be said to have become the first choice of major enterprises, and has run through all aspects of manufacturing such as design, production, management and service. Now artificial intelligence technology has been widely used in security, e-commerce, finance, medical care, education, personal assistant, autonomous driving and other fields. Artificial intelligence technology can automatically perform tasks, no longer need manual processes or tasks as in the past, improve enterprise performance and productivity, and can also exceed the human limit, give full play to the value of data and create huge benefits for enterprises.

At present, artificial intelligence has actually penetrated into all fields of our lives. In real life, almost everyone has access to AI artificial intelligence. From the aspects of food, clothing, housing and transportation to the new economy, until the meaning and value of life itself, artificial intelligence can be said to have brought rapid changes to various industries of human beings.

Wen | Chen Gen

In the era of a large change in the past 100 years, artificial intelligence serves as the core technology of the new round of technology revolution and the industry revolution. For the importance of global development, artificial intelligence technology determines that the international leadership of the big country is almost one inevitable. However, the uncertainty of artificial intelligence development has also exacerbated the difficulty of technical risk management, bringing new challenges for global governance.

First, the high productivity of artificial intelligence also means high fortification.Wealth accumulation and speed and international strength contrast will have a more obvious partial wild, that is, the rich country is rich, the strong country is more strong, and the poor country is the stronger, the weak country weakens, resulting in more wealth unevenness and unfair Incorrectization, more conflict with conflict and conflict and terrorism, this is a greater uncertainty and governance of global governance.

Second, artificially intelligent high execution also means high destructive power.If the new weapons and network viruses, there is an endless body condition that has the human body conditions and the unknown tired action, which will bring benefits to the implementation tasks of various countries, but may also be used by some forces, thus giving international Society has brought serious safety hazards, and may even bring disasters to the entire human society.

Finally, the high intelligence of artificial intelligence also means high political. Advanced artificial intelligence gives the technical owner’s outstanding additional advantages, robotic industry, gene sequencing, automatic driving, intelligent finance, smart city, big data processing, natural language processing, image identification, intelligent military system and other artificial intelligence, Change the country’s core competitiveness and economic and social and industrial structure, thereby changing the power structure in many fields.

Because of this, as a big country as the development and application of artificial intelligence, China and the United States play an extremely important role in artificial intelligence, and both parties have unique advantages that other countries cannot copy.How to prevent or reduce artificial intelligence technological advances on the negative impact of global sustainable development and strategic stability requires the work of China and the United States.

In fact, competition in the field of artificial intelligence is not an absolute zero and game, and there is a cooperative development and mutual benefit between parties. China is relatively leading in experimental research and results, and the United States is more leading in basic research and cutting-edge technology exploration. There is a broad cooperative space in both parties.

To this end, China and the United States should be in-depth thinking, work together to launch a formal dialogue on the application of artificial intelligence in safety and economic sectors.Promote the transparency of artificial intelligence research and development, driving its beneficial results to achieve reasonable allocation in a global scale, to minimize the competitive situation that may lead to catastrophic conflicts, and promote the formation of reasonable, benign competition and cooperative relations.

As Kissinger said, China and the United States are the most capable of impacting world progress and peace in technology, political experience and history, and solving the important issues in both parties in cooperation will be China and the United States. Common responsibility for peace and progress.