Group 1: Financing and Investment - Recently, Mianbi Intelligence successfully completed a new round of financing amounting to several hundred million yuan, led by Hongtai Fund, Guozhong Capital, Qingkong Jinxin, and Moutai Fund [1] - Since 2024, Mianbi Intelligence has maintained a steady financing pace, completing three rounds of financing, with significant investments from various venture capital firms [1] - Mianbi Intelligence and Zhipu AI are among the few companies that have successfully secured continuous financing amidst a challenging environment for large model companies [1] Group 2: Commercialization in Automotive Sector - Mianbi Intelligence is accelerating its commercialization process, particularly in the automotive industry, with the launch of the "Little Steel Cannon Super Assistant cpmGO," the world's first pure edge-side intelligent assistant for vehicles [2] - The debut of the MAZDA EZ-60, equipped with Mianbi's edge-side model, marks a new phase in the commercialization of edge-side large models in automotive cockpits [2] - The company has established partnerships with leading automotive manufacturers such as Changan Automobile, SAIC Volkswagen, and Great Wall Motors to promote large-scale deployment of edge-native intelligent cockpits [2] Group 3: Expansion into Vertical Industries - Mianbi Intelligence is also making strides in vertical sectors such as law and education, contributing to the development of national-level legal AI infrastructure [3] - The company has assisted in the launch of a judicial vertical large model that has supported over 291,000 cases and generated 11,600 draft documents since its trial run in January 2024 [3] - In education, Mianbi has partnered with Tsinghua University to introduce an AI learning assistant, aiming to create an automated classroom model with a student graduation rate exceeding 40% [3] Group 4: Technological Advancements - Mianbi Intelligence's edge-side model capabilities are continuously evolving, with the MiniCPM series gaining widespread recognition [4] - The MiniCPM-o 2.6 model features 8 billion parameters and supports innovative functionalities such as real-time interaction, achieving international leadership in image understanding and speech processing [4] - The MiniCPM series has surpassed 10 million downloads, establishing itself as a technical benchmark for edge-side large models globally [4]

Core Insights - Recently, Mianbi Intelligent completed a new round of financing amounting to several hundred million yuan, led by Hongtai Fund, Guozhong Capital, Qingkong Jinxin, and Moutai Fund, marking the third round of financing since 2024 [1][2] - Mianbi Intelligent has rapidly developed a complete matrix of full-modal, multi-modal, and foundational models, continuously pushing the boundaries of edge large model capabilities [1] - The MiniCPM series has achieved over 10 million downloads, recognized as the most downloaded and popular Chinese large model on Hugging Face in 2024 [1] Financing and Investment - The recent financing will further establish Mianbi Intelligent's efficient large model technology and product barriers, accelerating industry empowerment and ecological expansion [2] - The company aims to promote the large-scale application of "edge brains" across various industries by collaborating with upstream and downstream sectors [2] Product Development - In September 2024, Mianbi Intelligent released the MiniCPM 3.0 model, outperforming GPT-3.5 with 4 billion parameters [1] - The MiniCPM-V 0.6 model, launched in August 2024, achieved state-of-the-art results in single-image, multi-image, and video understanding with only 8 billion parameters, matching GPT-4V capabilities [1] - The first full-modal model, MiniCPM-O 2.6, was introduced in January 2025, enabling real-time interaction with 8 billion parameters [1] Applications and Collaborations - Mianbi Intelligent launched the "MiniCPM Super Assistant cpmGO," the world's first pure edge intelligent assistant for vehicles [2] - The company participated in the development of the "Faxin Legal Foundation Model," which has been released by the Supreme People's Court [2] - In collaboration with Tsinghua University, Mianbi Intelligent introduced the AI Student Growth Assistant "Qingxiaoda," providing personalized intelligent assistants for all undergraduate students [2]

vivo端侧大模型团队 投稿 量子位 | 公众号 QbitAI 在端侧设备上处理长文本常常面临计算和内存瓶颈。 vivo AI研究院 推出的EdgeInfinite算法专为端侧设备设计，让设备处理超长文本时更加高效流畅，该方法能够在不到10GB GPU内存的设 备上处理长达128K tokens的输入。 该研究成果已中稿ACL 2025。 以下是更多详细内容介绍。 EdgeInfinite:解决端侧设备长文本处理的高效算法 端侧LLM在实际应用中会遇到很多长文本输入的场景（例如通话摘要和个人文档总结），但由于端侧设备的资源限制，现有的LLM在部署到 端侧后都无法处理很长的上下文。 这是由于现在LLM都是基于Transformer架构，其计算耗时和内存占用会随着输入长度增加而显著增长，尤其当需要将Transformer类模型 部署到端侧设备上时，面临的挑战会愈发突出。 为了解决这类问题， vivo AI研究院 提出了一种用于端侧设备的长文本算法—— EdgeInfinite ，该算法通过一个可训练的 门控记忆模块 将记忆压缩算法集成到了 Transformer架构 中。 本方法与原生的Transformer架构 ...

Core Insights - The smartphone industry is undergoing an AI revolution, with manufacturers increasingly integrating AI features into their new products, marking a shift from traditional hardware innovation to AI-driven functionalities [2][5][14] - IDC forecasts a dramatic increase in AI smartphone shipments in China, with a year-on-year growth of 591% in 2024, and a penetration rate rising from 3% in 2023 to 22% [4] - The competition among smartphone manufacturers is shifting from hardware specifications to AI capabilities, emphasizing the need for end-to-end AI design from chips to operating systems [8][13] Group 1: Industry Trends - The AI smartphone market is expected to reach 1.18 billion units by 2025, accounting for 40.7% of the overall market [4] - High-end smartphones priced above $600 are projected to exceed 30.9% of the market share, with AI features contributing 75% of their premium pricing [4] - The average replacement cycle for smartphones has extended to 51 months, prompting manufacturers to focus on AI to drive consumer upgrades [5] Group 2: Technological Developments - The new generation of smartphones must feature advanced AI capabilities, including large model computing power, system-level AI integration, and proactive service in various scenarios [8][16] - AI's impact on imaging technology is significant, with innovations allowing for real-time analysis and optimization of images, enhancing capabilities beyond traditional photography [10][11] - The relationship between hardware manufacturers and AI developers is evolving, with companies like Qualcomm and Huawei creating ecosystems that support AI development and deployment [17][22] Group 3: Competitive Landscape - Major smartphone manufacturers are divided into three camps: Apple, Huawei, and an open ecosystem represented by brands like Xiaomi and Honor, each pursuing different AI strategies [20][22] - Huawei is positioned to lead in the AI smartphone market due to its strong R&D investment and technological capabilities in AI chipsets and cloud collaboration [22][23] - The future of smartphones may not solely rely on traditional devices, raising questions about the evolution of AI-native smart devices beyond current smartphones [23][24]

Core Insights - The article introduces Mixture-of-Lookup-Experts (MoLE), a new architecture designed to optimize the deployment of Mixture-of-Experts (MoE) models, particularly in resource-constrained environments [1][28] - MoLE addresses the challenges of high memory usage and transmission delays associated with traditional MoE during inference by replacing matrix operations with lookup tables [28] Group 1: MoLE Architecture - MoLE activates only a small subset of experts needed for each token during inference, significantly reducing computational load while maintaining a large parameter scale [1] - The architecture allows for the pre-computation of input-output mappings stored as lookup tables, enabling efficient retrieval during inference [5][6] Group 2: Training Phase Differences - In the training phase, MoLE modifies the input to routed experts from the previous layer's output to shallow embedding tokens, facilitating the pre-computation and storage of lookup tables [8] - MoLE employs an activation strategy that activates all routed experts during training, eliminating the need for sparse activation to control computational load [9] - The loss design in MoLE focuses solely on language modeling loss, without additional load balancing loss terms [10] Group 3: Inference Phase Process - During inference, MoLE constructs lookup tables from the embedding layer's weight matrix, allowing for direct retrieval of expert outputs based on token IDs [15] - The lookup table is stored in lower storage devices, and during inference, the corresponding expert outputs are retrieved and loaded into memory for computation [16] Group 4: Performance and Efficiency - MoLE's computational complexity during inference is comparable to dense models and traditional MoE models, while significantly reducing transmission overhead [17] - Experimental results indicate that MoLE achieves performance on par with MoE while drastically reducing transmission costs by over a thousand times [20][28] Group 5: Experimental Results - The experiments conducted on the Pile dataset show that MoLE maintains performance equivalent to MoE while using the same training parameters and inference activation parameters [20] - MoLE demonstrates lower inference latency compared to MoE, especially in batch decoding scenarios, highlighting its advantages in high-throughput tasks [28]

Core Viewpoint - The article highlights the rapid advancements in automotive AI technology, particularly focusing on the end-side large model developed by Mianbi Intelligent, which has achieved remarkable speed and efficiency in vehicle applications, revolutionizing the industry standards for AI integration in cars [1][14]. Group 1: Product Launch and Features - Mianbi Intelligent's cpmGO, a pure end-side large model-driven intelligent assistant, was showcased at the Shanghai Auto Show, marking a significant milestone in automotive AI [9][12]. - cpmGO boasts features such as 91% execution accuracy, local data processing, and robust performance in weak network conditions, making it a pioneering product in the industry [10][28]. - The product integrates multi-modal perception and interaction, allowing users to control vehicle functions through voice commands with high accuracy [30][31]. Group 2: Technological Innovations - The cpmGO model is powered by the MiniCPM, which operates entirely on the vehicle's local system, ensuring data privacy and rapid response times [27][28]. - The system's GUI Agent can understand and execute screen commands, enhancing user interaction by performing tasks autonomously based on context [33][36]. - The collaboration with major chip manufacturers like Qualcomm and Intel supports the optimization of cpmGO across various platforms, ensuring compatibility and performance [11][13]. Group 3: Industry Impact and Future Trends - The article discusses the shift in the automotive industry towards end-side AI models, which are less dependent on cloud services, addressing issues like latency and data security [38][42]. - The partnership between Mianbi Intelligent and Intel aims to redefine the next generation of in-vehicle AI systems, emphasizing the importance of local processing capabilities [40][48]. - The emergence of end-side models is seen as a response to the challenges of cloud-based solutions, positioning them as the future of automotive intelligence [44][46].