Microsoft, Google and Meta bet on synthetic data to build AI models
Every clever response from chatbots is supported by massive amounts of data — in some cases trillions of words need to be extracted from articles, books, and online reviews to teach artificial intelligence systems to understand users' queries. The industry's traditional view is that creating the next generation of artificial intelligence products will require more and more information. However, there is a big problem with this plan: the quality data that can be provided on the internet is limited. To get this data, the common practice of artificial intelligence companies is to either pay publishers millions of dollars to license content or download data from websites and expose themselves
Massive data (603138.SH): net loss of 81.1827 million yuan in 2023
Gelonghui, April 18 | Massive Data (603138.SH) released its 2023 annual report. During the reporting period, it achieved operating income of 262 million yuan, a year-on-year decrease of 13.57%; net profit attributable to shareholders of listed companies - 81.1827 million yuan; and basic earnings per share - 0.29 yuan.
Massive Data (603138.SH): The database product was not used in iFLYTEK's Spark model
Gelonghui, March 19 | Massive Data (603138.SH) said on the investor interactive platform that the company's database products were not used in iFLYTEK's Spark model.
Guojin Securities: Innovative drugs are an important field for the implementation of scientific and technological progress. Overall, performance will rise simultaneously with market performance
Guojin Securities released a research report saying that global innovative drugs are an important field for realizing scientific and technological progress, and China's innovative drug industry chain is on the rise. With the full restoration of in-hospital diagnosis and treatment, the introduction of national incentives and support policies for innovative drugs one after another, and the pace of internationalization of Chinese pharmaceutical companies, the industry as a whole will usher in simultaneous improvements in performance and market performance. It is recommended to invest in two types of targets: 1) choose a racetrack, preferring domestic targets mapped by metabolic, neurological, and self-immune indications and cutting-edge innovators in overseas breakthroughs; 2) select leaders, which are optimistic about bottom-up and internationalization breakthroughs. Recommended attention: Cinda Biotech (01801), Kangfang Biology
South Korea plans to invest 121.9 billion won this year to support the cloud computing industry
South Korea's Ministry of Science and ICT said on Monday that it plans to invest 121.9 billion won (about 91.5 million US dollars) this year to support South Korea's cloud computing industry, an increase of 17.2 billion won over last year. South Korea's Ministry of Science and Technology said that cloud technology plays a key role in advancing artificial intelligence (AI) because it provides high-performance computing capabilities and massive data storage capabilities. In terms of details, Korea plans to invest 24 billion won in the development and commercialization of innovative cloud services, 7 billion won in cloud-based software as a service (SaaS), and 80 billion won in the transformation of software to SaaS this year
Massive data (603138.SH): 2023 pre-loss of about 85.5 million yuan
Gelonghui, January 24 | Massive Data (603138.SH) announced the 2023 annual performance pre-loss announcement. According to preliminary estimates by the financial department, the company expects to achieve net profit attributable to shareholders of listed companies in 2023 of around -85.5 million yuan. Compared with the same period last year (legally disclosed data), there will be a loss. The company expects to achieve net profit attributable to shareholders of listed companies after deducting non-recurring profit and loss of about -96.1 million yuan in 2023.
Massive Data (603138.SH): Massive GaussDB database management system is a distributed database
Gelonghui, January 5 | Massive Data (603138.SH) said on the investor interactive platform that the company's massive GaussDB database management system is a distributed database.
Massive data (603138.SH): The company's core product database is a basic software
Gelonghui, January 5 | Massive Data (603138.SH) said on the investor interactive platform that the company is headquartered in Beijing, and the company's core product database is basic software.
Massive Data (603138.SH) granted 11.32572 million restricted shares at a price of 9.05 yuan/share
Massive Data (603138.SH) announced that restricted stocks stipulated in the company's 2023 restricted stock incentive plan...
Massive Data (603138.SH): The Vastbase Massive Database is mainly compatible with mainstream server operating systems such as Euler Operating System, Kirin Operating System, and Unicom Operating System
On December 14, Gelonghui Massive Data (603138.SH) stated on an interactive platform that currently the company's self-developed Vastbase massive database is mainly compatible with mainstream server operating systems such as Euler Operating System, Kirin Operating System, and Unified Communications Operating System.
Massive Data (603138.SH): The self-developed VastBase massive database is mainly compatible with mainstream domestic general computing power platforms such as Kunpeng CPUs, Feiteng CPUs, and Haiguang CPUs
Gelonghui, December 14|Massive Data (603138.SH) said on an interactive platform that currently the company's self-developed VastBase massive database is mainly compatible with mainstream domestic general computing power platforms such as Kunpeng CPUs, Feiteng CPUs, and Haiguang CPUs.
Soul announced the launch of the self-developed large-scale model “SoulX”, which can cover 100 types of fine-grained risk categories
Sina Technology News on the afternoon of December 11. Recently, SOUL, a new social networking platform, announced the launch of SoulX, a major self-developed language model. According to our understanding, the model is based on massive data training and has capabilities such as prompt driving, controlled generation of conditions, context understanding, and multi-modal understanding. While ensuring smooth, natural, and emotional conversations, it can also cover 100 types of fine-grained risk categories. A security system is constructed through strategies such as training data security screening, secure SFT data construction, RLHF security alignment, and inference interception, etc., to ensure the quality and safety of content production in large models. In the field of application
Amazon allegedly set up an AI model training team to spend millions
Sina Technology News reported on the evening of November 8, Beijing time, that people familiar with the matter revealed that Amazon will invest millions of dollars to train large-scale language models (LLM), hoping that it will be comparable to the top artificial models of OpenAI and Alphabet. According to sources, the internal code name of Amazon's artificial intelligence model is “Olympus,” and the model has 2 trillion parameters. This scale may make it one of the largest models currently being trained in the world. OpenAI's GPT-4 is currently one of the best models in the world. It is rumored that the model has 1 trillion parameters. Amazon declined
Massive data (603138.SH): Huakun Zhenyu is one of the company's ecological partners in the server field
Glonghui, Oct. 24 | Massive Data (603138.SH) said on the investor interactive platform that by the end of the third quarter of 2023, the company had completed compatibility and adaptation work with more than 1,000 software and hardware products from nearly 700 partners. Huakun Zhenyu is one of the company's ecological partners in the server field.
Massive data (603138.SH): net loss of 85.35 million yuan in the first three quarters
GLONGHUI, October 20: Massive Data (603138.SH) announced its third quarter report. Operating income was 163 million yuan, down 18.35% from the previous year, net loss was 85.35 million yuan, net loss was 85.35 million yuan, after deducting non-net loss of 92,829,800 yuan, and basic earnings per share of -0.301 yuan.
Ding Kuiling, President of Shanghai Jiaotong University: Sustainable scientific and technological development requires two strong foundations
Sina Technology News On the morning of September 7, at the 2023 Bund Conference, Ding Kuiling, an Academician of the Chinese Academy of Sciences and president of Shanghai Jiao Tong University, pointed out that only sustainable technology can create a sustainable future, and for science and technology to be sustainable, there are two main points: first, it is necessary to establish the two foundations of basic research and talent cultivation. Ding Kuiling pointed out that dingtian means continuously challenging cutting-edge scientific issues and card and own technical problems, striving for forward-looking basic research and key core technologies, and achieving major breakthroughs. For example, the Zhangjiang Institute of Advanced Studies at Shanghai Jiao Tong University continues to explore the storage of DNA information
Stealing the limelight by Nvidia, Qualcomm is also starting to develop generative AI
Sina Technology News reported on the evening of September 5, Beijing time. According to reports, due to the title of the world's largest chip company being taken away by Nvidia and the uncertainty of cooperation with Apple, Qualcomm is now also starting to develop generative artificial intelligence technology. Qualcomm's technology has penetrated almost every smartphone. It pioneered the ability to connect wirelessly in the 1980s until today's 5G modems (modems). Qualcomm charges patent fees for every device that uses its patented core technology to communicate. Now, Qualcomm is also studying a new kind of artificial intelligence using generative artificial intelligence (A
Massive data (603138.SH): Net loss of 59.024,000 yuan in the first half of the year
On August 15, | Massive Data (603138.SH) released its 2023 semi-annual report. The reporting period achieved operating income of 105 million yuan, a year-on-year decrease of 20.99%; net profit attributable to shareholders of listed companies - 590.204 million yuan; net profit attributable to shareholders of listed companies minus non-recurring profit and loss - 64.061 million yuan; basic earnings per share - 0.208 yuan.
Huawei Releases New AI Storage Product in the Era of Big Models Zhou Yuefeng: Data Storage Has Become a Key Infrastructure for Big AI Models
Sina Technology News reported on the afternoon of July 14 that Huawei today released a new AI storage product for the big model era, providing storage solutions for basic model training, industry model training, and segmented scenario model training and reasoning. In the process of developing and implementing large model applications, enterprises face four major challenges: first, data preparation time is long, data sources are scattered, and collection is slow; secondly, multi-modal models use massive text and images as training sets, and training set loading efficiency is low; third, large model parameters are frequently adjusted, and training platforms are unstable; finally, the implementation threshold for large models is high, system construction is complex, and resource scheduling is difficult. Huawei adapts to the big model
Massive data (603138.SH): Half year net profit advance loss of about 58.9 million yuan
On July 13, | Massive Data (603138.SH) announced the advance loss announcement for the 2023 semi-annual results. According to preliminary estimates by the financial department, the company expects net profit attributable to shareholders of listed companies to be around -58.9 million yuan in the half year 2023. Compared with the same period last year (statutory disclosure data), there will be losses. The company expects to achieve net profit of around -64 million yuan after deducting non-recurring profit and loss attributable to shareholders of listed companies in the half year of 2023. In the first half of 2023, the company maintained technological leadership and continuous innovation through continuous investment in R&D, and continued to improve utilization
No Data