Search AI Products and News

At 2025 China Internet Conference, Zhou Hongyi highlighted 3 AI security risks: 1) Hallucination may cause critical errors; 2) Lowers attack barriers; 3) Enhances state-level cyber threats. 360 launched "AI Security Expert" and "Model Guardian" for real-time defense.....

NVIDIA CEO Jensen Huang discussed AI's global impact at China's Chain Expo, highlighting 1.5M China-developed projects on NVIDIA platforms and Omniverse's role in digitizing supply chains. He predicted AI will drive industrial revolution.....

Zhiyuan Robotics secures strategic investment from CP Group, marking a new phase in Chinese robotics' global expansion. The partnership aims to enhance tech innovation and access international markets, focusing on life sciences and retail sectors.....

BAAI released RoboBrain2.0 (32B) and RoboOS2.0 framework. RoboBrain2.0 excels in spatiotemporal cognition and complex tasks, while RoboOS2.0 is the first embodied AI SaaS framework supporting lightweight deployment and multi-robot collaboration. Both are now open-source.....

The 2025 World AI Conference, themed 'Intelligent Era, Global Collaboration', will be held in Shanghai from July 26-28. It will launch an international AI open-source initiative and showcase latest AI technologies, building on its success since 2018 (300k+ visitors in 2024). China also plans a BRICS AI cooperation center.....

Ziyuan Robotics patented a training method for robot motion control models. It optimizes parameters by comparing actual and expected contact sequences, using swing excitation values. Expected sequences are derived from phase variables and gait mapping, improving motion precision and stability.....

OpenAI has officially announced the date and location of its next developer conference (DevDay), which will be held on October 6, 2025, in San Francisco. This conference is expected to attract more than 1,500 developers and is anticipated to be the largest developer event to date. The agenda for this DevDay will be rich and diverse, featuring multiple important sessions. The conference will include live-streamed keynote speeches, during which OpenAI will share its latest developments and future vision in the field of artificial intelligence. In addition, participants will also be able to

Tomorrow, the 2025FORCE Original Power Conference will be held in grand style. During the conference, the capability upgrade of the BeanPod Large Model family will be revealed. At the same time, the highly anticipated new BeanPod Video Generation Model will also be officially released. It is reported that this new model has multiple excellent features. In terms of multi-camera narration, through efficient model structure design, as well as multimodal positional encoding and multitask unified modeling technology, the model can seamlessly support unique and stable multi-camera expression, making video narration more smooth and natural.

In the field of artificial intelligence, the 'amnesia' phenomenon frequently encountered in long conversations significantly reduces user experience. To address this issue, the百家 AI team from Beijing University of Posts and Telecommunications has launched the first large model memory operating system -- MemoryOS. This system, through unique design, helps AI systems achieve persistent memory and personalized interaction, marking an important step forward for the development of AI agents. The inspiration for MemoryOS comes from both computer operating systems and the memory mechanisms of the human brain. It constructs...

At the 7th Beijing Zhiyuan Conference, Wang Xingxing, CEO of Unitree Technology, publicly stated that the ultimate goal the company has always adhered to is to make robots work - whether in household scenarios or industrial applications. Dancing and fighting are just training methods and technical verifications on the way to achieve this goal. We hope to empower humanoid robots through AI, enabling them to complete various full-body movements and eventually handle practical tasks such as serving tea and cooking meals. Wang Xingxing pointed out that currently, through embodied intelligence demonstrations such as fighting and dancing by Unitree Technology, they are actually training and testing robots' ability to...

At the BITA Conference held today in 2025, the Beijing Academy of Artificial Intelligence formally released the Wu World series of large models, comprehensively showcasing its latest research results and strategic layout in the direction of physical general artificial intelligence (AGI). The Wu World series aims to break through the boundary between virtual and real worlds and empower the physical world, covering four cutting-edge models: the multi-modal world model Emu3, brain science model Jianwei Brainμ, embodied intelligence brain RoboBrain 2.0, and microscopic life model OpenComplex2, constructing a framework from...

Recently, the Beijing Academy of Artificial Intelligence (BAEI), in collaboration with Shanghai Jiao Tong University and other institutions, officially released a new generation of ultra-long video understanding model - Video-XL-2. The introduction of this model represents a significant breakthrough in the field of open-source ultra-long video understanding technology, injecting new vitality into the development of multimodal large models for understanding long video content. In terms of technical architecture, Video-XL-2 mainly consists of three core components: a visual encoder, the Dynamic Token Synthesis (DTS) module, and a large language model (LLM). The model adopts

Programmers have a new tool to cheer about! LocAgent, a graph-indexed large language model (LLM) agent framework designed specifically for code localization, has been jointly developed by the research teams from OpenHands, Yale University, University of Southern California, and Stanford University. Excitingly, LocAgent's code localization accuracy reached 92.7%, and this new achievement will be officially presented at the 2025 ACL conference. Programmers often struggle with identifying the root cause of code problems in their work. For example

Recently, the Beijing Municipal Bureau of Economy and Information Technology officially issued the 'Action Plan for AI Empowered New Industrialization in Beijing (2025)'. The plan aims to promote the deep integration of artificial intelligence with industry to facilitate intelligent upgrades in manufacturing and enhance overall production efficiency. Image source notes: Images generated by AI; image authorization service provider is Midjourney. The primary task of the plan is to build high-quality industry datasets, encouraging manufacturing enterprises and research institutions to actively collect and process data, forming practical datasets, and addressing relevant requirements.

At the first International General Artificial Intelligence Conference TongAI, Alibaba Mum officially launched its LMA2 advertising large model series - the URM universal recall large model. This new model's debut in the field of generative recommendation (AIGR) marks another significant breakthrough for Alibaba Mum in intelligent advertising technology. The design concept of the URM universal recall large model is to accurately predict consumer shopping needs. Through deep learning and big data analysis, this model can effectively identify users' potential consumption intentions, thereby helping businesses.

Recently, Zhiyuan Robotics, a leading company in the field of embodied intelligence in China, announced that its new generation of humanoid robot product, Lingxi X2, has officially initiated partner recruitment, covering three models: Lingxi X2 (Interaction Edition), X2Pro (Exploration Edition), and X2Ultra (Flagship Edition). According to Zhiyuan Robotics' plan, the Lingxi X2 series robots will achieve large-scale delivery in the second half of 2025 and plans to reach the target of delivering thousands of units by the end of 2026. The core competitiveness of Lingxi X2 lies in its breakthrough emotion interaction capability.