Core Viewpoint - The article discusses the evolving landscape of artificial intelligence (AI) processing solutions, highlighting the need for companies to balance current performance with future adaptability in AI models and methods. Various processing units such as GPUs, ASICs, NPUs, and FPGAs are being utilized across different applications, from high-end smartphones to low-power edge devices [1][12]. Summary by Sections AI Processing Units - Companies are exploring a range of processing units for AI tasks, including GPUs, ASICs, NPUs, and DSPs, each with unique advantages and trade-offs in terms of power consumption, performance, flexibility, and cost [1][2]. - GPUs are favored in data centers for their scalability and flexibility, but their high power consumption limits their use in mobile devices [2]. - NPUs are optimized for AI tasks, offering low power and low latency, making them suitable for mobile and edge devices [2]. - ASICs provide the highest efficiency and performance for specific tasks but lack flexibility and have high development costs, making them ideal for large-scale, targeted deployments [3]. Custom Silicon - The trend towards custom silicon is growing, with major tech companies like NVIDIA, Microsoft, and Google investing in tailored chips to optimize performance for their specific software needs [4]. - Custom AI accelerators can provide significant advantages, but they require a robust ecosystem to support software development and deployment [4]. Flexibility and Adaptability - The rapid evolution of AI algorithms necessitates flexible hardware solutions that can adapt to new models and use cases, as traditional ASICs may struggle to keep pace with these changes [4][5]. - The need for adaptable architectures is emphasized, as AI capabilities may grow exponentially, putting pressure on decision-makers to choose the right processing solutions [4][5]. Role of DSPs and FPGAs - DSPs are increasingly being replaced or augmented by AI-specific processors, enhancing capabilities in areas like audio processing and motion detection [7]. - FPGAs are seen as a flexible alternative, allowing for algorithm updates without the need for complete hardware redesigns, thus combining the benefits of ASICs and general-purpose processors [8]. Edge Device Applications - Low-power edge devices are utilizing MCUs equipped with DSPs and NPUs to meet specific processing needs, differentiating them from high-performance mobile processors [10]. - The integration of AI capabilities into edge devices is becoming more prevalent, with companies developing specialized MCUs for machine learning and context-aware applications [10][11]. Conclusion - The edge computing landscape is characterized by a complex mix of specialized and general-purpose processors, with a trend towards customization and fine-tuning for specific workloads [12].

Group 1: Market Performance - The easing of Middle East conflict concerns has led to a rebound in U.S. stocks, with the Dow Jones up 0.75%, S&P 500 up 0.94%, and Nasdaq up 1.52% [1] - Chip stocks rose by 3%, with AMD increasing nearly 9%, while airline stocks also saw significant rebounds, with United Airlines up over 6% [1] - Chinese concept stocks outperformed the market, with the Chinese concept index rebounding over 2%, and RGC (Regenerative Medicine) soaring over 280% [1] Group 2: Economic Indicators - Following news of Iran's willingness to resume negotiations, the U.S. dollar index hit a daily low, while crude oil and gold prices experienced significant drops, with crude oil falling 10% from its daily high [1] - U.S. Treasury yields hit a daily low but later rebounded after a stable auction of 20-year Treasury bonds [2] Group 3: Industry Developments - The National Medical Products Administration of China is seeking public opinion on optimizing the review and approval process for innovative drug clinical trials, proposing to reduce the review period from 60 days to 30 days [4] - Guangdong Province has issued an action plan to accelerate the high-quality development of the nuclear medicine industry, aiming to cultivate 3-5 leading enterprises by 2030 [5] - Job demand for humanoid robots has surged by 409% year-on-year in the first five months of this year, according to Zhilian Recruitment [6] Group 4: Stock Market Insights - The Shanghai Composite Index fell below the critical support level of 3380 points but found support in the 3350-3360 range, while the ChiNext Index rebounded after hitting the 2030 support level [8] - The military industry sector has shown strong performance due to the Paris Air Show and geopolitical tensions, while the innovative drug sector has experienced fluctuations [8] Group 5: New Policies and Initiatives - Starting July 1, Dalian and Hubei provinces will implement a tax refund policy for overseas travelers, expanding the network of tax refund stores significantly [11] - The Chinese Ministry of Finance announced a reduction in the minimum purchase amount for tax refunds from 500 yuan to 200 yuan, enhancing the convenience of the refund process [11] Group 6: Company Announcements - Xiaomi is set to launch its new SUV model, the YU7, at the end of June, which emphasizes long battery life, intelligent driving, and ecosystem connectivity [12] - The company aims to replicate its success in the smartphone industry within the automotive sector by leveraging its large user base and innovative capabilities [12]

Group 1: Industry Insights - The AI energy crisis is prompting scientists to explore neuromorphic computing technology, aiming for efficiency similar to the human brain [1] - A newly developed supercomputer, occupying only two square meters, mimics the structure of the human brain and could operate 250,000 to 1,000,000 times faster than a biological brain while consuming only 10 kilowatts of power [1] - Neuromorphic computing is inspired by human brain functions, allowing neurons and synapses to work together effectively for data processing, and is seen as a key direction for next-generation AI [1] - The global neuromorphic computing market is projected to grow exponentially, reaching $1.81 billion by 2025 [1] Group 2: Company Developments - Keda Intelligent has partnered with Fudan University, focusing on brain-like intelligence research, including neuromorphic computing simulation and algorithms [2] - Xiangyu Medical has established the Sun-BCI Lab for brain science, developing EEG collection devices, brain-controlled products, and self-developed algorithms [2]

1—5月份，新建商品房销售面积35315万平方米，同比下降2.9%；其中住宅销售面积下降2.6%。新建商品房销售额34091亿元，下降3.8%；其 中住宅销售额下降2.8%。（国家统计局官网） 点击按钮▲立即预约 5月70城房价同比降幅继续收窄 6月16日，国家统计局公布数据显示，2025年5月份，70个大中城市中，各线城市商品住宅销售价格环比下降，同比降幅继续收窄。新房市场 中，一线城市新建商品住宅销售价格同比下降1.7%，降幅比上月收窄0.4%。二、三线城市新建商品住宅销售价格同比分别下降3.5%和4.9%，降 幅分别收窄0.4%和0.5%。二手房市场中，一线城市新建商品住宅销售价格同比下降1.7%，降幅比上月收窄0.4%。二、三线城市新建商品住宅销 售价格同比分别下降3.5%和4.9%，降幅分别收窄0.4%和0.5%。 6月16日消息，据国家统计局，5月份，社会消费品零售总额41326亿元，同比增长6.4%，比上月加快1.3个百分点，增速创2023年12月以来最 高。其中，除汽车以外的消费品零售额37316亿元，增长7.0%；5月社会消费品零售总额环比增长0.93%。消费品以旧换新政策持续显效，限额 以 ...

鹭羽 发自 凹非寺 量子位 | 公众号 QbitAI 「西部世界」真的要来了！科学家们正试图 为AI装上人类大脑 。 最新进展由美国国家实验室主导。科学家们正在试图将科幻拉进现实：打造一台占地仅两平方米、神经元数量堪比人脑皮层的超级计算机。 更令人惊叹的是，计算表明，这台神经形态计算机的运行速度可能比生物大脑快 25万到100万倍 ，而功耗仅需 10千瓦 （仅略高于家用空调 的能耗） ，这无疑是对当前AI发展困境的一剂强心剂。 然而相比之下，自然界最强大的智能体——人类大脑，每天只需消耗约 20瓦 ，仅相当于家用LED灯泡的功率。科学家们不禁思考： 能否让 AI也像人脑一样高效？ 神经形态计算：向大脑学习 在人类大脑中，约有860亿个复杂神经元相互工作，并通过100万亿个突触共同构建起一张巨大的信号传递网络。 神经形态计算受其结构和功能启发，采用模仿生物神经网络的节能型电子和光子网络，即 脉冲神经网络 （SNN） 构建，旨在将记忆、处理 和学习整合到一个统一的设计中。 答案是： 神经形态计算 。 这项旨在模拟人脑结构和运作方式的前沿技术，正被视为下一代AI的关键方向，其核心目标之一，就是用"灯泡级"的能耗驱动 ...

鹭羽 发自 凹非寺 量子位 | 公众号 QbitAI 「西部世界」真的要来了！科学家们正试图 为AI装上人类大脑 。 然而相比之下，自然界最强大的智能体——人类大脑，每天只需消耗约 20瓦 ，仅相当于家用LED灯泡的功率。科学家们不禁思考： 能否让 AI也像人脑一样高效？ 答案是： 神经形态计算 。 这项旨在模拟人脑结构和运作方式的前沿技术，正被视为下一代AI的关键方向，其核心目标之一，就是用"灯泡级"的能耗驱动强大的智能。 最新进展由美国国家实验室主导。科学家们正在试图将科幻拉进现实：打造一台占地仅两平方米、神经元数量堪比人脑皮层的超级计算机。 更令人惊叹的是，计算表明，这台神经形态计算机的运行速度可能比生物大脑快 25万到100万倍 ，而功耗仅需 10千瓦 （仅略高于家用空调 的能耗） ，这无疑是对当前AI发展困境的一剂强心剂。 目前人工智能正面临一场"能源危机"，随着大语言模型等技术的爆炸式发展，其惊人的耗电量已成为无法忽视的沉重负担。 预测显示，到2027年，仅运行这些模型的电费就可能高达25万亿美元——甚至超过美国当年的GDP。 神经形态计算受其结构和功能启发，采用模仿生物神经网络的节能型电子和光子 ...

Core Insights - A team from Tokyo University of Science has developed a self-powered artificial synapse with high color discrimination ability, closely approaching human eye capabilities, marking a significant advancement in machine vision research [1][2] - The rapid development of artificial intelligence has increased the demands on machine vision, which traditionally consumes substantial power, storage, and computational resources [1] - The new artificial synapse system integrates two types of dye-sensitized solar cells, enabling it to convert solar energy directly for power, making it suitable for energy-efficient edge computing applications [2] Group 1 - The artificial synapse can distinguish colors within the visible spectrum at a resolution of 10 nanometers, comparable to human vision [2] - It exhibits bipolar response characteristics, generating positive voltage under blue light and negative voltage under red light, allowing it to perform complex logical operations typically requiring multiple traditional optoelectronic components [2] - The system demonstrated an 82% accuracy rate in classifying up to 18 combinations of colors and human actions using a single device [2] Group 2 - This technology has the potential to provide human-like vision capabilities to everyday devices, with broad application prospects in autonomous driving, medical health devices, and consumer electronics [2]

如果您希望可以时常见面，欢迎标星收藏哦~ 来源：内容 编译自 eenewseurope 。 荷兰的 Innatera 公司推出了第一款采用神经形态架构的商用微控制器，适用于传感器应用。 Pulsar 芯片采用异构架构，将模拟和数字神经形态模块与传统的卷积神经网络加速器和 RISC-V 内核相结合。与传统 AI 处理器相比，这款芯片的延迟降低了 100 倍，功耗降低了 500 倍，芯片 尺寸为 2.6 x 2.8 毫米，采用台积电标准 28 纳米工艺制造，批量成本不到 5 美元。 nnatera 联合创始人兼首席执行官 Sumeet Kumar 向eeNews Europe表示："Pulsar 不仅仅是一款 AI 芯片——这是世界上第一款量产的神经形态微控制器，代表着我们将智能带到边缘的根本性转 变。" 去年传感器出货量达 380 亿个，到 2030 年将增长至 600 亿个，所有这些传感器生成数据的速度 都将快于我们将其发送到云端的速度。边缘处理将不再是可选的。但目前部署在微控制器上的模型 有限，应用程序开发人员必须在功能、精度和功耗之间做出权衡。 此次发布是我们十余年神经形态计算领域深入研究和工程的巅峰之作 ...

28款车型齐聚，两款全球首秀 "这届上海车展上感觉有一半人都是国际友人。" 4月23日，时值2025（第21届）上海国际汽车工业展览会首日，一些参展的媒体人如此感慨。 展会上国际面孔的增多，一方面说明着中国的汽车开始更多被全球关注，另一方面，也是国际品牌们更 加重视中国市场的具体表现。其中，梅赛德斯-奔驰就以本次车展为契机，大秀新平台、新产品、新技 术，俨然将上海车展办成了奔驰的春晚。 从去年以来，针对竞争愈加激烈的中国市场，国际品牌们纷纷交出自己的应对答案，而这一次的上海车 展上的众多大招，便是奔驰应对中国市场变局、继续深度连接中国的答案。 本次上海车展，奔驰携旗下梅赛德斯-奔驰、梅赛德斯-AMG、梅赛德斯-迈巴赫和G级越野车全品牌共28 款车型亮相。 其中，MMA平台的首款国产车型——全新梅赛德斯-奔驰纯电长轴距CLA，以及面向未来的顶级豪华 MPV车型Vision V概念车在本次车展上完成全球首秀，全新梅赛德斯-AMG GT 63 4MATIC+正式上市。 全新纯电长轴距CLA车型是奔驰针对中国市场智能与豪华变革之下的答案之作。 在奔驰的体系中，这款车是迄今为止最智能的奔驰。新车搭载了奔驰自研全新架构M ...

如果您希望可以时常见面，欢迎标星收藏哦~ 来源：内容来自 dqindia ，谢谢。 新加坡国立大学的研究人员证明，单个晶体管可以模拟神经和突触行为，使受大脑启发的计算更接 近现实。 这是通过将块体终端的电阻调整到特定值来实现的，这可以控制晶体管中发生的两种物理现象：穿 透碰撞电离和电荷捕获。此外，该团队还构建了一个能够在神经元或突触状态下运行的双晶体管单 元，研究人员将其称为"神经突触随机存取存储器"，简称 NS-RAM。 世界上最复杂的计算机已经存在于我们的大脑中。研究表明，人类大脑总体上比电子处理器更节 能，这要归功于近 900 亿个神经元，它们彼此之间形成了大约 100 万亿个连接，而突触会随着时 间的推移调整其强度——这一过程被称为突触可塑性，它是学习和记忆的基础。 几十年来，科学家一直试图利用人工神经网络 (ANN) 复制这种效率。ANN 近来推动了人工智能 (AI) 的显著进步，其灵感大致来源于大脑处理信息的方式。但是，虽然它们借用了生物学术语，但 相似之处只是表面的——基于软件的 ANN，例如那些为 ChatGPT 等大型语言模型提供支持的 ANN，对计算资源有着极大的需求，因此也需要电力。这使 ...