苹果 AI 端侧详细解读 20240707_原文 2024年07月07日 23:42 发言人 00:00 能理解对吧？或者说之前的这么一个说，这是关于这个用户体验反馈的第一点。第二点的话是他为什么是 北美的英语版本会优先放开这个事情？对于北美的，对，首先是对于这个语言，他现在训练比较完善或者 比较完整的。主要是英文版本，像中文或者是其他的几个大语种，包括像西班牙语或者是法语前都还没有 开，或者说没有完全准备好这个相关的其他语言的这些模型的一个训练，或者相关的这些法律法规，各方 面的一些问题的支持。这个是还需要花一点时间去再去更新和升级的那第二个的话是那个地区，它现在主 要是在于北美的英语版本，那为什么不放开其他地方的英语版本的这些体验？这个跟服务器的建设相关。 目前就你们的跟踪来看的话，苹果在推理测的那个服务器建设上面，主要的建设地区的话还是在它本部北 美地区为主。像中国地区或者是或者是欧洲地区，相关的苹果的自己的私有云服务器的建设，其实都还没 有完全的去搭建或者准备好啊。这个也是为什么是会限制在北美应用版本体验的核心原因。 发言人 01:23 在这个推理服务器上面，我们直接就进入到下一个观察的节点，就是苹果 ...

您的线上直播间然后我是今天的主持人斯奇那今天我们也是来到了上海世界人工智能大会的展会现场然后呃站在我身边的是我们的呃半导体传媒以及呃半导体传媒还有呃 其他一些产品的一些在管的经理罗英宇罗总欢迎我们罗总来到我们这个直播间带我们走进这个世界人工智能大会我们还是先请罗总跟镜头前的观众打个招呼吧各位线上的投资者大家下午好非常高兴跟各位一起来逛一逛全球人工智能大会 好 那我们今天的话其实我们本次的2024年的这个世界人工智能大会呢其实也是创了很多历史的新高然后也是有很多吸引人的地方那我们也是很好奇它里面到底有哪些新奇的一些展品展示我们也是希望罗总能够带领我们去参观这个展会那我们也是想先请罗总就是给我们介绍一下就是我们这一次的大会其实有哪些的亮点和看点呢 因为我看到还有一些什么特斯拉呀还有我们的一些微软同讯啊其实他们在这里都会有相应的一些展区我们很好奇这次的亮点有哪些去年其实我也 我们也在投资一起参观过上一届的全球人工智能大会 那么对比上一届来说的话我能感受得到这一届的人是参观的人是非常的多这个大家可以感受得到就现场这个人流是非常的允许对对对人挤人的这样一个状态所以也多少能够反映就是当前大家对于人工智能的这样的一个产业的 ...

Hello, investors. It's been a year since we last met. In order to promote the basic knowledge of ETF and improve our ability to operate in real-time, this year's ETF Handling Contest will be officially launched on July 1st. In addition to handling in real-time, starting from July, we will have a live broadcast with the fund every week. We will discuss the investment value of long-term hot spots and indexes with you. Today is the first live stream after the tournament. We are honored to have two bigwigs from ...

高盛：日本电子元器件和半导体目标价和盈利预测调整增加对受益股的关注和可能的业绩超预期 ...

AI 科技评论：您认为当前机器学习和深度学习领域面临的最大挑战是什么？您的研究 如何帮助应对这些挑战？ 赖载兴: 最大的挑战可能是可解释性，这样模型就可以用来完成艰巨的任务。我们研究 中使用的混沌边缘原理可以作为提高人工智能可解释性的理论基础。 AI 科技评论：您能否介绍一下本研究的应用前景和潜在挑战？ 赖载兴: 我们的研究是概念验证的第一步，表明利用复杂系统科学中的原理可以帮助创 建更好的人工智能模型。从长远来看，它可以带来更好、更复杂的训练算法或模型架 构。然而，由于大型语言模型等最先进的模型非常庞大且复杂，因此实现更好的人工 智能模型非常具有挑战性。 Q: 您如何看待量子信息科学和复杂系统研究在未来技术发展中的作用？他们将如何推 动跨学科创新？ 赖载兴: 量子信息科学正沿着一条不可阻挡的轨迹迅速发展，这不仅得益于它巨大的潜 力，更源于它对量子技术革新的驱动作用。随着我们对量子世界的认识日益加深，对 通信、计算和安全等领域的高级功能需求不断增长，探索量子领域已成为一个合乎逻 辑的下一步。制造量子设备是一个复杂的过程，它要求精密的工程技巧和跨学科知识 的融合。量子信息科学有潜力成为连接多个前沿创新的关键纽带 ...

Hello, investors. It's been a year since we last met. In order to promote the basic knowledge of ETF and improve our ability to operate in real-time, this year's ETF Investment Competition will be officially launched on July 1st. In addition to the real-time operation, we will have a live broadcast with the fund every week from July. We will discuss the investment value of long-term hot spots and decryption indexes with you. This is the first live stream after the tournament. We are honored to have two expe ...

注公众号:永木纪要 2 July 2024 | 6:33AM KST South Korea Technology: Semiconductors - Memory Favorable Al/server demand and limited supply to lead to stronger memory pricing; raise TP, Buy Hynix/SEC We expect a stronger memory (DRAM and NAND) pricing trend for the remainder of the year as well as for next year, with our updated industry S/D model indicating a larger undersupply especially for DRAM. Memory demand trends are mixed with server-related demand seeing continued strength from Al and a recovering trend from ...

Summary of Conference Call Notes Industry or Company Involved - Focus on Apple's battery supply chain opportunities [1] Core Points and Arguments - Discussion on Apple's AI potential and its impact on the battery supply chain [1] - Previous in-depth reports on individual stocks related to Apple's battery and AI sectors were mentioned [1] - The call aimed to address key concerns regarding Apple's AI and its implications for the battery industry [1] Other Important but Possibly Overlooked Content - The timing of the call was strategically chosen for a Sunday morning to provide insights [1] - Emphasis on collecting core concerns from stakeholders regarding Apple's AI and battery supply chain [1]

Adynatrace ©2023 Dynatrace 2024 年 AI 发展状况 更多一手调研纪要和研报数据加V：shuinu9870 采用 AI 面临的挑战和组织成功的关键策略 更多资料加入知识星球：水木调研纪要 关注公众号：水木纪要 • 74% 的组织已因 AI 而调整所招聘的职位和技能 • 24% 的组织计划在未来 12 个月内因 AI 而调整所招聘的职位和技能 增强 AI 提供的解决方案的可解释性 78% �dynatrace 用作模型训练和提示工程输入的数据质量 94% 自动智能优化 AI 提示的能力 88% 团队技能或培训 84% 提高 AI 摄取的信息的准确性和质量 82% 瑞典 • 76% 的组织已因 AI 而调整所招聘的职位和技能 • 20% 的组织计划在未来 12 个月内因 AI 而调整所招聘的职位和技能 消除偏见 76% Glynatrace 2024 年 AI 发展状况 | 36 更多资料加入知识星球：水木调研纪要 关注公众号：水木纪要 比荷卢经济联盟 样本库：50 名受访者（荷兰 32 名、比利时 10 名、卢森堡 8 名） 技术领导者期望 AI 实现的主要改进包括： • 94% ...

Financial Data and Key Indicators Changes - The company expects a significant increase in net profit for the first half of the year, projecting a range of ¥0.97 to ¥1.12 billion, representing a year-on-year growth of 50.62% to 73.91% [12] - The company's revenue for 2022 was ¥5,336 million, with a projected increase to ¥6,228 million in 2023, reflecting a year-on-year growth of 16.7% [26] - The net profit for 2023 is expected to be ¥372 million, a substantial increase from a loss of ¥262 million in 2022 [26] Business Line Data and Key Indicators Changes - The company has established two major technology bases: "HuaJia Mos Metro Smart Brain" and "IDPS Urban Traffic Brain," focusing on smart rail transportation and smart city governance [1][32] - The company has released the first AI model for urban traffic, which is expected to enhance existing business scenarios and create a comprehensive AI strategy [19][17] Market Data and Key Indicators Changes - The smart transportation market in China is projected to exceed ¥400 billion in 2023, with significant growth expected in the coming years [3][15] - The company is positioned to benefit from the expanding smart city market, which is expected to reach ¥33 trillion by 2024 [3] Company Strategy and Development Direction - The company is focusing on enhancing its AI capabilities and has developed multiple AI models, including natural language and visual models, to support its smart transportation initiatives [33] - The strategic partnership with Huawei aims to deepen collaboration in smart rail and smart city sectors, enhancing the company's technological capabilities [31] Management's Comments on Operating Environment and Future Outlook - Management expressed optimism about the growth potential in the smart transportation sector, driven by AI technology and the increasing demand for smart city solutions [19][30] - The company acknowledges the competitive landscape but believes its established technology bases and partnerships will provide a strong advantage [19] Other Important Information - The company has filed 126 new patents and received 82 patent grants in 2023, indicating a robust intellectual property portfolio [27] - The company is also expanding its VR and AR capabilities, positioning itself as a leader in the audio module supply chain for these technologies [20][24] Q&A Session Summary Question: What is the expected growth in net profit for the upcoming years? - The company anticipates net profits of ¥0.83 billion, ¥1.93 billion, and ¥3.09 billion for 2024, 2025, and 2026, respectively, with corresponding P/E ratios of 95.8, 41.3, and 25.8 [29] Question: How does the company plan to leverage AI in its operations? - The company plans to utilize its AI models to enhance traffic management and customer service in smart transportation, aiming for a comprehensive AI strategy [19][17]