360 Smart Life held a large visual model and AI hardware new product launch conference, announcing its official entry into the SMB market.

On May 31, the 360 ​​Smart Life Visual Model and AI hardware new product launch conference was held in Beijing. With the theme of "Security Protection, Wisdom for the Future", the press conference launched the 360 ​​Intelligent Brain and Vision large-scale model and a variety of new AI hardware products, and announced that 360 Smart Life has officially entered the SMB market, realizing a large-scale application from home scenes to enterprise commercial use. step across.

Not long ago, Zhou Hongyi, the founder of 360 Group, requested in a letter to all employees to "fully embrace artificial intelligence" and "leave homework" for all product managers and engineers of the company to reshape the entire product line with large-scale model capabilities. In April this year, the browser and search products supported by 360 Intelligent Brain were opened for internal testing. Today’s press conference is a new answer sheet handed over by 360 Intelligent Life.

In recent years, with the development of artificial intelligence technology, the limitations of traditional deep learning algorithms in security scenarios have become increasingly prominent. The algorithm is universal and the scenarios are not universal. It cannot solve the problem of scene fragmentation at a low marginal cost. At the same time, problems such as limited edge algorithms and limited memory also limit the further popularization and implementation of AI. In contrast, most The model has better generalization capabilities, is expected to further break through the accuracy and data limitations of traditional algorithms, and can also solve the data dependence problem of traditional deep learning algorithms. On this basis, it can not only reduce training and development costs, but also implement vertical scenarios. A quick landing. It can well make up for the limitations of traditional deep learning algorithms.

Currently, the core of multimodal capability enhancement is mainly the cognition, reasoning, and decision-making capabilities of large language models. Therefore, 360 Smart Life combines visual perception capabilities with the 360 ​​Intelligent Brain large language model, and fine-tunes industry data for security scenarios to create a professional visual and multi-modal large model - 360 Intelligent Brain·Visual large model.

According to Sun Hao, general manager of the 360 ​​Vision Cloud business line, the 360 ​​Intelligent Brain Vision model currently focuses on three capabilities: open object detection (OVD), image title generation, and visual question answering (VQA). Among them, open target detection is based on learning a large amount of image and text data on the Internet, allowing the model to comprehensively learn more target categories; image title generation is an advanced form of understanding pictures, which can satisfy users to quickly and accurately obtain information. needs; visual question and answer is based on the recognition, understanding, and reasoning of pictures to answer various questions raised using natural language. These three capabilities can be widely used in long-tail target detection, store inspection, nursing, equipment inspection and other scenarios, and can form a good complement to traditional deep learning algorithms.

However, the implementation of security scenarios still needs to rely on the collaboration of large and small models and the integration of devices, edges, and clouds. Among them, the small model can be used as a sample value judgment to help the large model learn and converge quickly; the large model can quickly transfer capabilities to the small model based on knowledge distillation, which can help the small model quickly generate and label samples to improve efficiency.

In terms of the application of AI, 360 Smart Life has been exploring for many years. Last year, it launched an intelligent security SaaS platform with scenario-based AI as the core - 360 Vision Cloud. 360 Vision Cloud not only undertakes the implementation of AI algorithms and subsequent large-scale visual models, but also provides a full set of intelligent security solutions for small, medium and micro enterprises. At present, 360 core intelligent security products have been connected to 360 Vision Cloud and can be seamlessly connected in the same scene to form a data visualization solution. Among them, AI LAB can effectively solve the scene adaptation problem of edge and end algorithms. The 360 ​​Vision Cloud, equipped with the 360 ​​Intelligent Brain·Vision large model, not only has 50 scene algorithms to choose from, but also can implement containerization of algorithm running, regularization of scene adaptation, and visualization of skills and detection results, which is especially suitable for small, medium and micro enterprises. Digital transformation and security capability building.

As we all know, small, medium and micro enterprises face problems such as excessive investment, difficult to measure output, and lack of security during the digital transformation process. The 360 ​​Vision Cloud originally has the advantages of low delivery threshold, low configuration method, low technical difficulty, low operation difficulty, and low usage cost. After equipped with the 360 ​​Intelligent Brain·Vision large model, it can accelerate the promotion of AI inclusive hardware products. Upgrade, let security products truly enter the civilian intelligent era.

At this conference, 360 Smart Life also released the 360 ​​Outdoor Ball Machine 6 Pro, the 360 ​​Dazzling Outdoor Gun and Ball All-in-one Machine, and the 360 ​​Binocular Splicing Panorama with the theme of "Inclusive AI, Renewal and Upgrade" New products such as cameras and 360 Vision Smart NVR have achieved full coverage of scene channels from indoor to outdoor, from home to commercial, and from online to offline. Empowered by the 360 ​​Intelligent Brain·Vision large model, these products have stronger edge computing capabilities and can achieve wider compatibility to adapt to different application scenarios.

From the release of 360 Intelligent Brain Vision large-scale model to the full-line upgrade of security products, 360 Smart Life has taken a key step in the era of AI large models, and officially entered the SMB market with fully upgraded 360 Smart Life products. In the future, 360 Smart Life will continue to delve into the forefront of intelligent security, explore the application of AI large models in security scenarios, and use inclusive AI to empower the digital upgrade of small, medium and micro enterprises.

At this press conference, Zhou Hongyi said that intelligent hardware will become the next outlet for the development of large models, and large models will bring intelligent hardware into the AIoT era. It is based on this judgment that Wang Rulin, President of 360 Smart Life, unveiled a new corporate brand concept: Safety·Intelligence. It also announced the launch of the "See a Better Life" charity plan, from "Guarding for Love" to "Safety and Wisdom", highlighting the dual renewal of 360 Smart Life at the strategic and brand levels. Security is born out of the gene and soul of 360 Smart Life, while Zhijian highlights the future strategic direction of "visual AI". In the future, 360 Smart Life will bear in mind the responsibility of sharing development with society, and is committed to protecting every family and every enterprise in the world, with visual AI security as the core, so that more people can see a better life.

The above is the detailed content of 360 Smart Life held a large visual model and AI hardware new product launch conference, announcing its official entry into the SMB market.. For more information, please follow other related articles on the PHP Chinese website!