2023 WAIC: Over 30 large models are key to their external appearance, application, and landing

Tech 2023-07-10 09:13:41 Source: Network

On July 8th, the three-day World Artificial Intelligence WAIC Conference closed in Shanghai. The exhibition area of this conference exceeds 50000 square meters, and a total of 133 forums were held

On July 8th, the three-day World Artificial Intelligence WAIC Conference closed in Shanghai. The exhibition area of this conference exceeds 50000 square meters, and a total of 133 forums were held. The conference reached a total of 210 upstream and downstream enterprises, reached an intention to purchase 11 billion yuan, and promoted the signing of 32 major industrial projects, with a total investment of 28.8 billion yuan.

It is worth noting that the AI big model has become the absolute focus of this conference. According to incomplete statistics, more than 30 large model platforms have been released and made public, with 60% of offline booths showcasing relevant introductions and applications of generative AI technology. 80% of the attendees' discussions revolve around large models.

From the research and development of generative AI technology, the appearance of large model platforms, to the layout of scene applications, many followers have also raised questions: has the large model application entered a mature landing period?

More than 30 large models made public

At the exhibition, Alibaba Cloud, Shangtang, Tencent, JD and other enterprises all released the progress of their big models with the help of the WAIC platform.

Among them, Alibaba Cloud announced that the AI painting creation model Tongyi Wanxiang opened a targeted test invitation. It is understood that Tongyi Wanxiang is based on the combined Generative model Composer developed by Ali, which can provide a highly controllable and free image generation effect by disassembling and combining image design elements such as color matching, layout and style. Only one model can support multi category image generation task capabilities. This research achievement has been published on the International AI Summit ICML2023.

At the same time, Alibaba Cloud Cloud has launched the big model assistant Magic GPT, which can call all big models with one click. At present, Magic GPT can now call the APIs of more than ten core AI models in the Magic Community. In the future, with the addition of more model APIs, the ability of Magic GPT will continue to enhance.

Tencent Cloud has publicly announced upgrades in three major areas: technology base, innovation scenarios, and AI for Science, which is another update following the industry's big model technology solutions in June. Among them, the newly upgraded Tencent Cloud's self-developed Star Pulse High-performance computing network can increase the GPU utilization by 40%, save 30%~60% of model training costs, and bring 10 times of communication performance improvement to the AI large model.

Shangtang Foreign Technology has launched a comprehensive upgrade of the "Shangtang Nissin SenseNova" large model system in multiple aspects. As a Natural language processing model with hundreds of billions of parameters, Shangtang discussed SenseChat 2.0 version broke through the limit of input length of large language model, and launched model versions with different parameter levels. The model parameters of SenseMirage 3.0, a self-developed and generative large model of Shangtang, have increased from 1 billion since its first release in April this year to 7 billion, enabling professional photography level image detail depiction.

In addition to leading companies in the field of internet technology, AI startups such as the Fourth Paradigm and Yunzhisheng have also competed to showcase their latest achievements in big models. At the same time, the academic community has also begun to conduct research on large models. ChatGLM-130B, a 100 billion parameter Chinese English dialogue model in the Knowledge Engineering Laboratory of the Computer Department of Tsinghua University, and MOSS in the Natural language processing Laboratory of Fudan University attracted the attention of participants.

As for the development of AI technology, Pan Xingang, the first author of DragGAN and assistant professor of the School of Computer Science and Engineering of Nanyang Technological University, mentioned at the forum that AI is becoming the basis for broader creative applications.

The implementation and application of large models have entered an acceleration period

Xu Li, chairman and CEO of SenseTime, said during the conference: "The breakthrough in the big model has set off a new round of technological revolution in artificial intelligence, followed by an explosive growth in industrial demand, and new application scenarios and models are emerging rapidly."

This viewpoint has also been validated on forums and exhibition booths.

Ant Group and Tsinghua University jointly released AI security detection platform "Ant Awareness 2.0". Ant Detection 2.0 can detect generative AI models using generative AI capabilities, identifying hundreds of risks in three major categories: data security, content security, and technological ethics, covering multiple data and task types such as tables, text, and images.

ZhongAn released the first AIGC application map of the insurance industry on the conference forum, and brought Zhong'an AIGC midrange Lingxi and the first batch of insurance vertical scenario AIGC application tools - Yi Chuang content operation platform and intelligent business analysis platform. Compared with the traditional "AI+insurance" application, Tupu conducts a feasibility study on the enabling value and future application potential of AIGC in 19 categories and 78 subcategories of business scenarios from six dimensions of product design and innovation, marketing and development, underwriting and claims settlement, customer service, smart office and R&D management, covering property insurance, life insurance and other businesses, involving agents, proprietary and other Marketing channel, Provide reference and inspiration for the digital transformation of the insurance industry.

At the NetEase booth, multiple artificial intelligence large model landing products of NetEase made their debut, showcasing their technical achievements in the artificial intelligence infrastructure layer, engine layer, platform layer, model layer, and application layer, including the unmanned loader developed jointly by NetEase Fuxi and China Construction Eighth Bureau, and the Chinese model Danqing self-developed by NetEase Fuxi. Among them, the latest application achievement of "Confucius", the first domestic education model developed by Youdao of NetEase, the Virtual humans speaking coach, also made its debut. NetEase staff disclosed that the application results of the "Confucius" model series, including Virtual humans speaking coaches, will be disclosed in detail at the end of July.

In terms of industry applications, different enterprises have similar perspectives: the introduction of large models is bound to enter industry applications. Zhou Jingren, CTO of Alibaba Cloud Intelligent Group, said that the release of Tongyi Wanxiang is a key step for Alibaba Cloud's big model to fully master multimodal capabilities, which will be gradually open to industry customers. He believes that "multimodality is an inevitable path for the evolution of large models, and we hope to use the capabilities of different modalities to serve thousands of industries and help enterprises fully embrace the era of intelligence

SenseTime hopes to continue to promote the leap forward improvement of AI infrastructure capability through "big model+big device", not only to create a more powerful basic model with general capabilities, but also to further integrate expertise in different vertical fields efficiently, and build a more knowledgeable and specialized professional model.

On the basis of the big model, Tencent has also provided its own answer on how to serve the development of vertical industries. Tang Daosheng, Senior Executive Vice President of Tencent Group and CEO of Cloud and Smart Industry Group, stated at the conference that building one's own exclusive model based on the industry model may be a better option for enterprises. At present, Tencent's own enterprise level applications have been integrated into the industry's large model capabilities.

How far is the true application?

The latest data from KPMG shows that as China vigorously promotes the construction of computing infrastructure, the scale of computing related industries in China can reach 24 trillion yuan.

The application of large models in the industry is expected to bring more development opportunities, but the challenge still lies in how to truly implement the application of large models. Yang Zhilin, founder of AI startup Moonshot AI and assistant professor at Tsinghua University, discussed his philosophy of innovation and achievement transformation. He said that there are still many unresolved issues with large models. When thinking about them, it is more important not to rely on single point thinking, but to systematically abstract the underlying commonalities of these problems and return to a more fundamental level to solve them.

In this regard, Jiang Jijun, chief technology officer of ZhongAn, said that with the implementation of more large vertical industry models and self-developed models, AIGC will show the potential of integration and migration. Through the combination of scenarios and data, AIGC will achieve personalized customization, risk control and other advantages, and bring more efficient, intelligent and humanized services.

Ant Group believes that the deep application of AI not only puts forward higher requirements for data, algorithms and computing power, but also poses more challenges to security, privacy and ethics. Only by ensuring data security and privacy protection, as well as sound artificial intelligence ethics and security, can AI technology truly unleash its application value.

Wu Yunsheng, Vice President of Tencent Cloud, Head of Tencent Cloud Intelligence, and Head of Youtu Laboratory, believes that the development of AI big model technology and industrial exploration cannot be separated from industrial chain collaboration and ecological co construction.

According to the research conducted by KPMG, the development of China's manufacturing industry is worth looking forward to. Zhang Qingjie, managing partner of KPMG China Digital Empowerment, said: "China's manufacturing industry has the characteristics of perfect industrial system, long industrial chain, and many industrial clusters, which makes the manufacturing industry very rich in digital scenarios. In terms of Digital transformation, it is expected to fully realize economies of scale and economies of scope. With the occurrence of systematic industrial reform, '5G+industrial Internet', 'million enterprises on the cloud', and 'integration of two technologies' will drive the comprehensive expansion of intelligent manufacturing scenarios, and the manufacturing industry is becoming a' universal intelligence 'computing power The largest potential market

Disclaimer: The content of this article is sourced from the internet. The copyright of the text, images, and other materials belongs to the original author. The platform reprints the materials for the purpose of conveying more information. The content of the article is for reference and learning only, and should not be used for commercial purposes. If it infringes on your legitimate rights and interests, please contact us promptly and we will handle it as soon as possible! We respect copyright and are committed to protecting it. Thank you for sharing.(Email:[email protected])

Mobile advertising space rental

Tag: 2023 WAIC Over large models are key to their