标签归档 上海龙凤论坛

Tesla investors’ day talks about the concept of "three self" products defined by automobile robots

At Tesla’s first investor day, Musk once again made a surprising statement: "Tesla cars are robots on wheels, and no one and no company can approach Tesla’s level in real-world AI." As soon as this was said, the two major topics of AI and robotics once again triggered industry discussions.
At the beginning of 2023, other new forces, such as ideals, also said that they should build systematic capabilities around AI robots. "Intelligent electric vehicles with autonomous driving will also become the earliest artificial intelligence robots."
From the perspective of the whole automobile industry, it is the concentration that first puts forward the advanced concept of "automobile robot". As early as the beginning of its establishment in 2021, Jidu took "automobile robot" as its product positioning, and automobile robot has also become another product masterpiece under the Geely SEA framework.
Objectively speaking, Musk’s statement made the world see the tremendous changes that AI technology has brought to the automobile industry, and also strengthened the development trend of "robot" of automobiles.

"Automobile robot" conforms to the future development trend, and "three self" is the basic attribute.

"No AI, no intelligence" has become the consensus of the automobile industry. In the past, fossil energy was the core power source of automobiles in the era of fuel vehicles 1.0, and electric vehicles 2.0 were driven by electricity. In the current era of smart cars 3.0, the underlying driving force came from AI technology.

Like Musk, Xia Yiping, CEO of Jidu, always emphasizes that AI technology is the "soul" when he mentions "car robot". Xia Yiping once said, "The era of smart car 3.0 is the era of automobile robots. The starting point of the era change is that the’ driving right’ of automobiles is transferred from human beings to AI, and AI drives the evolution of automobiles." Concentration also defines the basic attributes of "automobile robot", that is, "free movement, natural communication and self-growth", which has also become the product concept of concentration automobile robot.
Why do you want to divide it like this? This is actually due to the nature that automobile robots are both cars and robots. When a car has the brain of AI, it is no longer just a simple means of transportation, but a "robot" with self-thinking and powerful ability. Therefore, automobile robots must first be able to "move freely" and rely on high-order intelligent driving to drive to where they need to go; Secondly, we should have enough high-level intelligent interaction ability, constantly understand the new needs of users, and complete effective communication and action execution; Finally, it can continue to grow through the learning of big data, understand users more and more, and improve its ability boundary.
From this point of view, "automobile robot" does not have to look like a human body, but whether it has the "three-self" ability. Even if it becomes human, it still can’t escape this underlying logic. Just like Tesla’s humanoid robot, it must be able to walk on its own, communicate smoothly, and learn and evolve continuously.

In addition to the standard demonstration of concentration, have car companies gone out of another new path? At present, not yet.

Whether Tesla said "cars are robots on wheels" or "artificial intelligence robots" that he wanted to build, or even independent brands directly label their products with "car robots", in fact, they are all different expressions of the same logic.

From hardware-led to software-defined, automobile is actually a transfer from automobile attributes to robot attributes. The advancement of intelligence must not be achieved overnight, and it needs a long-term accumulation process. Although everyone is called a "robot", their abilities may differ by a hundred thousand miles.

Compared with Tesla and other followers, Jidu, as the pioneer of new species of automobile robots, has integrated AI genes into the product concept of "three self".

Geely SEA’s vast architecture is fully redundant in terms of space, electricity, intelligence, autonomous driving, safety and performance. Therefore, the centralized automobile robot also has a considerable degree of freedom, which can integrate higher-order intelligent capabilities and further improve the "three-self" level of automobile robots.

Specific to the function, the centralized high-level intelligent driving is empowered by Baidu Apollo’s L4-level automatic driving technology, and one of the core functions, point-to-point navigation assistance PPA, can meet the "free movement" in most daily use scenarios. High-level intelligent interaction is supported by off-line voice, touch, visual recognition and other multi-touch interactions to realize "natural communication" in any scene. Among them, the global off-line voice interaction recognition speed is 500ms, the response speed is within 700ms, and it can be used smoothly even when the network is disconnected. In addition, based on the underlying technical support accumulated by Baidu AI in the past 10 years, Jidu automobile robot can continuously learn and "grow up" according to user behavior habits and driving data.

Although the birth of automobile robot is defined by software, it is equally excellent in mechanical quality. As Jidu automobile robot is another masterpiece of Geely SEA’s vast architecture, it is "of the same ancestry" with Volvo, Polar Star and Lotus in terms of mechanical performance and safety guarantee, and the quality of the whole vehicle is directed at international luxury brands.
In addition, Geely’s manufacturing technology has also reached the international first-class level. For example, Geely Hangzhou Bay Smart Factory integrates advanced manufacturing technologies such as 5G, AI and industrial big data, and the four production workshops are highly automated. As the first batch of "future factories" in Zhejiang Province, Geely Hangzhou Bay Smart Factory has become a "new model" for China’s intellectual creation.

It is conceivable that the product quality of Jidu automobile robot created by Geely’s excellent craftsmanship is worthy of recognition.

In the real version of Iron Man and Jarvis, Jidu automobile robot will take the lead in applying the China version of ChatGPT.

If Jarvis, the smart and reliable AI housekeeper in Iron Man, depicts the ultimate form of artificial intelligence in the future for us. Then, with the gradual maturity of AI empowerment and core technologies, automotive robots are turning Jarvis in the movie into the reality of the automotive industry.
ChatGPT-like application of Baidu Wenxin is a turning point. It can be said that with ChatGPT, AI’s empowerment for the automotive industry has evolved.

As an AI conversation model, ChatGPT has a natural connection with cars, which has the potential imagination to subvert the voice assistant in the intelligent cockpit.

On February 14th this year, Jidu announced that it would integrate the comprehensive capabilities of Baidu Wenxin to create the world’s first large-scale artificial intelligence interactive experience for smart car scenes, so as to support automotive robots to achieve further progress in natural communication. As the first "crab-eating" brand, Jidu’s move has also brought unprecedented changes to smart cars.
The reason why Jiji was able to announce that it was the first to realize the first launch of ChatGPT-like technology is also because the pure-blood automobile robot of Geely SEA has already possessed high-level intelligent capability. A smarter smart car can access more powerful AI capabilities, which is a condition that traditional cars do not have. Simply put, a tractor can’t match a Boeing engine.

As we all know, Jidu is the first car company to introduce Qualcomm 8295 into the cockpit, and it is also the first to localize the AI algorithm, which can ensure the efficient response speed of cockpit interaction even in the scene of poor network environment or even disconnection.

At present, the voice interaction of most car companies can not be separated from the dependence on the network, and there is still a distance from "natural communication".

At present, the pre-embedding of hardware by Jidu automobile robot in advance, coupled with the empowerment of Wen Xin in a word, means that the product concept of "natural communication" will be pushed to a new height.

In the face of new technology, both traditional car companies and new forces are struggling to seize the opportunity to seize the commanding heights of the industry. With the continuous empowerment of AI technology, the application of intelligent interaction not only enhances the brand image and experience, but also accelerates the digital transformation process of the automotive industry, laying the foundation for the advanced automotive robots.
With all parties in the industrial chain, including car companies and Internet giants, the pace from technology research and development to commercialization is unprecedented. Geely SEA’s vast architecture of integrated "automobile robots" is also being imitated by more and more industry partners and gradually "integrated".

Artificial intelligence replaces manual labor, uses machine learning to carry out scientific research and match the best alloy!

A large group of researchers found that machine learning can be used to help metallurgists find the best metal mixture to make the required alloys. In this way, they found an alloy with a thermal coefficient lower than that currently recorded.

The research was carried out by scientists from Max Planck Institute for Iron Research, who cooperated with colleagues from darmstadt University of Technology, Delft University of Technology and Royal Institute of Technology.

For thousands of years, people have been mixing metals according to their own needs and learned a lot about alloy manufacturing. But finding the right combination always requires some inspiration, patience and luck.

Therefore, most alloys are made from a single base metal (such as iron), and the obtained characteristics are observed by adding a small amount of other metals.

However, in the past few decades, the situation has begun to change-some researchers have begun to make alloys containing several metals in equal parts.

Of course, it is much more difficult to make this alloy with specified characteristics. In this new work, researchers apply machine learning to help this process. They first reduced the test space to only one application-making alloys that will not expand and contract too much when exposed to temperature changes.

In order to create a machine learning application, researchers searched and found metal features that can be used for comparison purposes, and then trained their systems using the information in the currently available databases. In the process, they developed a process to find an alloy suitable for the required goal.

The team’s process was reduced to three main steps: First, they created a new mixture using a model based on information stored in a database describing metal characteristics. Then, they use the second model to help predict the properties of some alloys they created in the first step. The last step is to study the candidate alloys produced by the system and select some of them for practical testing.

Using this system, the researchers obtained 1000 candidates, which were narrowed down to three alloys. Then, they made three kinds of alloys using the mixture described in the system and tested their physical properties.

The team used data from real alloys to train the system and repeated the whole process. They chased it for seven times and found an alloy with a thermal coefficient lower than the current record.

In their paper published in the journal Science, the team described their three-step process and its performance in the test. Scientists from the Institute of Metals, Chinese Academy of Sciences published an article on the prospect in the same issue, describing the work done by the team on this new project.

Reference: https://techxplore.com/news/2022-10-machine-optimal-mixture-metals-desired.html.

JD.COM Group released Q4 and annual financial report: jingdong cloud laid out industrial AI to promote artificial intelligence landing industry.

JD.COM Group (NASDAQ: JD, HKEx: 9618) released its fourth quarter and annual results in 2022, with annual net income exceeding one trillion yuan. In the fourth quarter, its net service income was 57.8 billion yuan (about 8.4 billion US dollars), a year-on-year increase of 40.3%. Jingdong cloud, as the core brand of technology and services provided by JD.COM, continues to exert its cutting-edge technology and digital infrastructure to boost the real economy to achieve high-quality growth with the ability of digital intelligence supply chain.

Since the comprehensive transformation to technology in early 2017, JD.COM system has invested nearly 100 billion yuan in technology, continuously strengthened its own technical capabilities and industrial digitalization capabilities, and provided technology and services to the outside world with jingdong cloud as its core brand. In the field of AIGC and big model, jingdong cloud’s artificial intelligence application platform of Yanxi is preparing for the industrial version of ChatGPT to accelerate the landing of artificial intelligence technology in the industry. At present, Yanxi virtual anchor has served more than 4,000 brands; JD.COM Exploration Institute upgraded the model of the Weaver Girl, and topped the list of SuperGLUE, an international authoritative task evaluation for complex language understanding.

Jingdong cloud continued to dig deep into cutting-edge technology, and the Weaver Girl model topped the national list again.

In the basic model of common language understanding, JD.COM Exploration Institute upgraded the Vega v2 model with larger scale, stronger performance and better mobility. This model can be widely used in many downstream natural language processing tasks, such as sentiment analysis, semantic matching, grammar correction, intelligent question answering, common sense reasoning and so on. On the SuperGLUE list of international authoritative complex language understanding task evaluation, Vega v2 model surpassed top international institutions such as Google, Microsoft, Facebook and OpenAI, and topped the world with an average score of 91.3.

This is not the first time that the Weaver Girl model has won the championship. Vega v1 ranked first in the top test GLUE in the global natural language processing field with a total average score of 91.3. Vega-MT also won 7 track titles in WMT2022 International Machine Translation Evaluation. The winning of Vega series models proves that JD.COM Exploration Institute’s multilingual natural language processing technology is leading in the field of super deep learning.

Jingdong cloud exports industrial AI capabilities, and digital intelligence technology helps thousands of industries to grow with high quality.

Jingdong cloud’s artificial intelligence application platform is preparing for the industrial version of ChatGPT, and has announced the "125" plan of landing application roadmap. JD.COM will focus on two advantageous industries, namely retail and finance, and carry out technical research around five types of applications: content generation, man-machine dialogue, user intention understanding, information extraction and emotion classification, so as to accelerate the development and landing of artificial intelligence technology in China with the strength of industrial AI and promote the development of real economy.

In the field of digital people, jingdong cloud has launched the virtual anchor of Yanxi. Driven by AI, the virtual anchor of Yanxi has a changeable image, a voice comparable to that of a real person, and a wealth of e-commerce knowledge accumulation. At present, it has interacted smoothly with the audience in hundreds of brand live broadcast rooms, widely serving 3C home appliances, beauty cosmetics, maternal and child, pets, home and other stores, and providing digital anchor services for many well-known brands such as Lenovo, Nut and Yili, bringing millions of GMV growth every day.

Jingdong cloud, as a "more industry-aware" cloud, will not only deepen its technology and maintain its leading position in the industry, but also rely on its own practical experience in the industry to continuously precipitate technology into services and land in major industrial scenes, so as to build an efficient driving force for thousands of industries to build a digital and intelligent supply chain, promote the construction of a modern industrial system and help the industry rebuild its global competitiveness.