从传说到故事，从3D影视到VR应用，人类创造虚拟世界的愿望是原始且永恒的。元宇宙作为一个既平行又独立于现实世界的虚拟空间，是一个影射现实世界、越来越接近真实的数字虚拟世界。同时，元宇宙不是完全虚拟世界，而是基于现实的虚拟世界；元宇宙不是某个虚拟空间，而是某个数字生活价值大于物理生活的奇点时刻；元宇宙与星辰大海不是对立的，而是加速人类进入星辰大海；元宇宙不是以游戏为主，而是虚实相融的数字文明；元宇宙不是5~10年后才成熟，而是有初级、中级、高级的不同阶段。
From legend to story, from 3D visions to VR applications, the desire of humankind to create a virtual world is primitive and eternal. As a virtual space that is parallel and independent of the real world, the metacosystem is a virtual world that insinuates the real world and is closer to the real world.
从ICT技术的视角看，元宇宙是互联网的下一个阶段，是用AR、VR、区块链、云计算、人工智能、数字孪生、3D、物联网、人机交互、数字货币、Web3.0等技术支持的虚拟现实的网络世界，而构成元宇宙的数字基础底座的核心是联接、计算和人工智能。
From the perspective of ICT technology, the meta-cosm is the next phase of the Internet, a virtual reality of the network with technical support from AR, VR, block chains, cloud computing, artificial intelligence, digital twin, 3D, object networking, human interaction, digital currency, Web3.0 and so forth, while the core of the digital base of the meta-cosm is connection, computing and artificial intelligence .
从连接人到连接物，从万物互联到万物智联，从现实到虚拟，网络需求永无止境。华为《智能世界2030》预测，到2030年，全球联接数将达到2000亿；人均无线蜂窝网络流量将超过600GB，增长40倍；XR的用户数将超过10亿；千兆及以上家庭宽带用户渗透率将达到55%，增长50倍；家庭月均网络流量将达到1.3TB，增长8倍。物理世界与虚拟世界的交互将达到一个全新的高度，网络的特征将表现为：交互网络、通感一体、算力网络、确定性体验、认知网络、立体超宽等，超现实的体验将无处不在。
From connecting people to connects, from connecting everything to connecting intelligence, from reality to virtual, the need for networks is endless. China’s Smart World 2030 projects that the number of global connections will reach 200 billion by 2030; per capita wireless cellular network traffic will exceed 600 GBs and increase 40 times; the number of users of XRs will exceed 1 billion; the penetration rate of broadband users in households in gigabytes and above will reach 55 per cent, or 50 times; and household average monthly network traffic will reach 1.3 TB, or eight times. The physical world interacts with the virtual world at a completely new height, and the network will be characterized by: interactive networks, perceptiveness, numeric networks, definitive experiences, cognitive networks, stereo-wideness, etc., and super-real experiences will be everywhere.
1. 交互网络：下一代网络是人机交互网络，提供以人为中心的超现实体验，从过去的人主动适配机器转变为机器主动适配人。在物理世界，网络将能够对语言、触觉、光感、脑机等多模信息进行采集；在虚拟世界，系统能够对3D视觉、触觉、听觉等多模交互进行呈现，从而实现对视觉、听觉、触觉、甚至嗅觉、味觉的传输。
1. Interactive networks: The next generation of networks is an interactive network of humans, providing a people-centred super-real experience, moving from a man-to-man-to-man to a machine-to-man-to-man. In the physical world, the network will be able to collect multi-model information such as language, tact, light, brain machines, etc.; in the virtual world, the system will be able to interact with 3D visuals, tactiles, hearing, tactiles, even smelling, taste, etc.
2. 通感一体：通信网络的感知能力正在越来越多地被发掘出来，通过通信与感知融合，网络将为平行世界提供现实复刻，利用虚拟技术对现实情况做出反应，为现实生活提供便利。无线的感知精度正在从亚米级发展到厘米级、毫米级，并与雷达、毫米波等技术配合，实现超灵敏多源融合感知与定位，为网联车、无人机、无人安防、环境监测、气象预报、医学影像、工业检车等提供超高精度定位。
2. Sensitivity integration: The communications network is increasingly being developed and, through communication and sensory integration, the network will provide a realistic re-entry into the parallel world, using virtual technology to respond to reality and facilitate real life. Wireless sensory accuracy is being developed from the Yami level to the centimetre level, the millimeter level, and in conjunction with such technologies as radar, millimetre waves, super-sensitive multi-source integration and positioning, providing super-high-precision positioning for network vehicles, drones, unmanned, environmental monitoring, meteorological forecasting, medical imagery, industrial inspection vehicles, etc.
华为创新性地在光纤通信网络中增加低成本超窄线宽激光器，光纤就具备智能感知能力，实现对震动、磁、电、热、声、光、气压等的感知能力，再结合人工智能技术，光纤通信能为智慧城市、智慧工厂、智慧园区等的应用场景提供高准确率的识别能力。Wi-Fi结合AI技术，可以识别老人摔倒、检测呼吸和心率、识别情绪，从而为健康检测提供服务，Wi-Fi与红外、雷达等结合，为交互游戏提供姿势、手势识别，为周界安防提供融合入侵检测等能力。
With the innovative addition of low-cost ultra-wire-wide lasers to fibre-optic communication networks, fibre-optics have the intellectual ability to sense vibrations, magnetics, electricity, heat, sound, light, air pressure, etc., combined with artificial intelligence technologies, fibre-optic communications can provide high-precision recognition for applications such as smart cities, smart factories, smart parks, etc. Wi-Fi combines AI techniques to identify older persons falling, detect respiratory and heart rates, recognize emotions, provide health testing, Wi-Fi combines with infrared, radar, etc., to provide position, hand recognition for interactive games, to provide cross-sectional security with capabilities such as integrating intrusion detection.
3. 算力网络：过去，网络提供电话和短信的通讯服务。如今，网络提供视频、IoT等内容服务，流量以下行为主，这些服务面向人的认知，需求特征是100ms时延和30fps的刷新频率。未来，元宇宙和数字孪生世界中，网络将以提供机器视觉、全息社会等为主要应用服务，数据流量以上行为主，服务处理更多从云端走向边缘，需求特征是10ms时延和120fps的刷新频率，网络将与高效算力、海量数据、泛在智能融合一体，构建融合的算力网络。
3. Arithmetic networks: used to provide telephone and text message communication services. Today, the network provides content services such as video, IOT, and traffics the following behavioural masters, which are oriented towards people's perceptions and are characterized by a need for 100 ms time delay and 30fps for refreshing frequencies. In the future, the network will provide a major application service in the meta-cosm and digital twin world, providing machine vision, a holo-social society, etc., with behaviour masters above data flow, and more services processing from cloud edges. The demand feature is a new frequency of 10 ms time delay and 120fps, which will be integrated with efficient computing, big data, general intelligence, and build an integrated computing network.
4. 确定性体验：过去，网络的特征是尽力而为，可以满足个人上网、企业办公等的日常需求。未来，虚拟世界中，无论是游戏、VR业务，还是数字孪生带来的城市治理、企业生产等场景，都需要确定性体验，网络需要以其灵活性和确定性，来应对未来业务的不确定性。华为用端到端切片的方式为垂直行业打造更加适配的逻辑“专网”和服务，并通过集群DC和边缘计算的布局，提供三级时延圈，也就是城市内1ms超低时延、城市群10ms低时延、骨干100ms中高时延。专网具备99.999%的高可靠性，满足行业生产控制系统要求，使能企业全要素上云。
4. The certainty experience: used to characterize the network as trying to meet the day-to-day needs of individuals to access the Internet, business offices, etc.
5. 认知网络：机械技术延伸了人类的身体，例如工具是手的延伸，汽车是脚的延伸，电子信息技术延伸了人类的神经系统，人类的各种连接使地球成为类人脑组织。从早期的电报到模拟电话、卫星通信、数字电话、移动通信、光通信、数据通信，再到物联网和未来的孪生网络，信息通信网络正在具备越来越强的认知能力，正在赋能人类在元宇宙世界进化为“高等生命”。
5. Cognitive networks: mechanical techniques extend the human body, such as tools are hand extensions, cars are foot extensions, electronic information technologies extend the human nervous system, and human connections make the Earth a cerebral tissue. From early telegrams to analogue telephones, satellite communications, digital telephones, mobile communications, light communications, data communications, and then to physical networking and future twin networks, information and communication networks are becoming increasingly cognitive and empowering humans to evolve into &ldquao in the meta-cosm world; high life & rdquao;
6. 立体超宽：个人、家庭、组织经历语音时代、百兆时代，进入千兆时代，面向3D立体的虚拟元宇宙世界，则需要万兆接入方式，从而支撑现实世界与虚拟世界的无缝对接。Sub100GHz支持万兆个人接入，200G-PON支持家庭万兆接入，毫米波WiFi支持组织万兆接入。无线、光纤、低轨卫星的组合实现空天地一体的无缝立体连续覆盖，为地面、高空航空器等应用场景提供时间与空间维度连续宽带体验。接入域、城域、骨干网、数据中心网络正在全面进入全万兆时代。
6. Stereo-wideness: individuals, families, tissues through the speech age, the megabytes, into the gigabytes, a virtual meta-cosmos world for 3D stereos, requires a megabytes of access to support the seamless interface between the real and the virtual world. Sub100GHz supports megabytes of personal access, 200G-PON supports family megabytes of access, and millimetre waves of WiFi supports organization megabytes of access. WiFi provides continuous coverage of the combination of wireless, fibre-optic, low-orbiting satellites, providing continuous broadband experience in time and space dimensions for applications such as ground-based, high-altitude aircraft.
因此，从联接百亿人到联接千亿物，面向元宇宙世界的通信网络将具备六大特征：立体超宽、确定性体验、智能原生、通信感知融合、安全可信和绿色低碳。
Thus, from connecting 10 billion people to connecting hundreds of billions of objects, a communications network to the meta-cosmos will have six main features: a stereo-broadness, a definitive experience, an intelligent birth, a sense of communication integration, safety and credibility, and a green low carbon.
超级算力是未来智能元宇宙世界的关键需求，算力与网络、AI等技术的结合，为智能世界提供无处不在的感知、永无止境的探索、无所不及的AI和源源不断的创新。
Super arithmetic is the key need of the future smart-constellation universe, combining arithmetic with technologies such as networks, AI and so forth, providing an ubiquitous sense of the intelligent world, endless exploration, ubiquitous AI and continuous innovation.
- 无所不在的感知：超级算力为个人和家庭提供基于AR/VR的超现实体验，为企业、工厂、农业等提供更纵深的感知；
- 永无止境的探索：超级算力帮助企业更精准地模拟现实，例如风洞仿真、新药探索，生态监测、海洋预测、气象和地震预测等探索；
- 无所不及的AI：超级算力为城市和出行提供更聪明的AI，如无人驾驶、智慧交通、智慧城市等；为医疗和教育提供更普惠的AI，如精准医疗、药物筛选、个性化教育等。
- 源源不断的创新：超级算力帮助企业挖掘数据价值，提供数据驱动的业务创新；帮助企业更精细化的资源使用，实现更高的运营效率。
今天，一部智能手机的算力，已经远远超过了当年阿波罗登月时主控计算机的能力。数字世界和物理世界无缝融合，人与机器实现感知、情感的交互，以及无所不及的AI，都在帮助人类获得超越自我的思想力与行动力，成为科学家的显微镜与望远镜，并让我们的认知跨越微小的夸克触及广袤的宇宙，使得千行万业从数字化走向智能化。
Today, the computing power of a smart phone is far greater than the ability of the computer in control at the time of the month of Apollo. The seamless integration of the digital and physical worlds, the interaction of human beings with machines, and the omnipotence of AI are helping human beings to gain a mind and action that transcends self, become the microscopes and telescopes of scientists, and bring our perception across tiny quarks to the vast universe, leading to a digitalization of industries into intelligence.
华为《智能世界2030》预测，到2030年，人类将进入YB数据时代，全球通用算力，将达到3.3 ZFLOPS，增长10倍；AI算力将超过105 ZFLOPS，增长500倍。未来计算将面临物理极限的挑战，不仅要从软件、架构和系统层面去创新，更要联合产业界共同探索新的计算基础，突破半导体物理层极限，构建智能、绿色、安全的未来计算。
By 2030, China's Smart World 2030 predicts that humanity will enter the YB data age, with global universal computing reaching 3.3 ZFLOPS, a 10-fold increase; and that AI will grow more than 105 ZFLOPS, a 500-fold increase. Future calculations will face the challenge of physical limits, not only in terms of innovation at the software, architecture, and system levels, but also in terms of joint efforts by industry to explore new computing bases, breaking the semiconductor physical layer limits, and building smart, green, and safe future calculations.
ICT技术的竞争领域正延续到元宇宙世界。所谓根深才能叶茂，中国需要坚持自主化的基础软硬件创新，构建持续发展的基石，使能上层应用生态百花齐放，推动数字经济高质量发展。
The area of competition in ICT technology is continuing into the Woncosmos world. So-called deep-seated talent, China needs to adhere to autonomous innovation in basic software and hardware, to build the foundations for sustained development, to allow the application of the upper layers of ecology to flourish, and to promote high-quality development in the digital economy.
在通用计算和AI计算领域，华为重点聚焦5大根技术的创新和研究，包括通用计算、AI计算、基础软件、终端操作系统、AI开发框架等。
In the area of universal computing and AI computing, China focuses on innovations and research in five major technologies, including universal computing, AI computing, basic software, terminal operating systems, AI development frameworks, etc.
① 通用计算：基于鲲鹏处理器构建全栈IT基础设施、行业应用及服务。
1 Universal calculation: complete storage of IT infrastructure, industry applications and services based on a thong processor.
② AI计算：基于昇腾系列处理器构建的全栈AI计算基础设施及应用。
2 AI Calculating: Acid computing infrastructure and applications based on totem series processors.
③ 基础软件：华为将自研多年并在电信领域规模商用的操作系统OpenEuler和数据库OpenGauss，全面开源出来，发展基础软件。
3 Basic software: China will develop basic software by fully generating OpenEuler and Database OpenGauss, an operating system that will be self-researched for many years and on a commercial scale in the field of telecommunications.
④ 华为OpenHarmony是智能终端操作系统，并且OpenHarmony和OpenEuler和实现了内核技术共享，搭载欧拉的设备可以自动识别和连接OpenHarmony终端，更好地服务数字全场景。在工业场景中，OpenEuler和OpenHarmony可以有更多协同，OpenEuler适用于高可靠性、强确定性的工业设备；而OpenHarmony适用于强交互性的工业终端。
OpenHarmony is an intelligent terminal operating system, and OpenHarmony and OpenEuler share in internal nuclear technology, with equipment on board Ora automatically identifying and connecting the OpenHarmony terminal and better serving the digital landscape. In the industrial scene, OpenEuler and OpenHarmony can have more synergy, with OpenEuler applying to high reliability, high certainty industrial equipment; and OpenHarmony applying to highly interactive industrial terminals.
⑤ 华为分层解耦的AI开发框架，为不同层级的AI开发者提供了开发工具与平台。AI开源框架昇思MindSpore，原生支持大模型，帮助开发者快速进行模型的开发与训练。昇腾应用使能MindX，快速帮助开发者完成AI应用的开发，实现极简AI开发。对于算子开发者，可直接使用异构计算架构CANN进行开发，打造良好性能。
5 China provides development tools and platforms for different levels of AI developers as part of the AI development framework for decoupling layers. The AI open-source framework, thinking MindSpore, supports large models and helps developers to develop and train models quickly. The application enables the developers to quickly complete the development of AI applications and achieves a very simple AI development. For arithmetic developers, the asymmetric architecture can be developed directly using CANN to build good performance.
元宇宙世界离不开人工智能和基于人工智能的AI模型。基于AI的计算中心能够提供大算力，加上大数据，可以使能大模型，孵化新应用。大模型天然具有平台属性，基于预训练大模型，开发者只需要做少量场景适配，就可以快速开发出高精度行业AI模型。大模型是促进AI技术生态的关键，华为工业化的AI开发新模式，正在促进AI行业发展。
The metaspace world cannot be separated from AI models based on artificial intelligence and artificial intelligence. A calculator based on AI can provide great arithmetic, coupled with big data, that can enable large models to hatch new applications. Large models naturally have platform properties, based on pre-training large models, and developers can quickly develop high-precision industry AI models with only a small number of fit-out scenarios. Large models are key to promoting AI's technological ecology, and China's industrialization AI develops new models and is contributing to the development of the AI industry.
- 鹏城实验室发布了两千亿稠密参数的中文NLP大模型“鹏程·盘古”；
- 鹏城实验室发布了面向生物信息研究的“鹏程·神农”；
- 基于武汉人工智能计算中心，武汉大学发布了面向遥感影像智能解译AI框架“武汉·Luojia”；
- 中国科学院自动化所发布了图文音三模态大模型“紫东·太初”。未来还会基于各地的特色孵化出更多领域的大模型。
在多模态大模型-跨模态检索与生成方面，目前已有成功实例。
Successful examples of multi-model large models -- cross-model retrieval and generation are available.
在以文搜图方面，模型达到了较高的精度，输入的文本和检索返回的图像结果的语义非常相关，输入“桌子上的蓝色小碗里盛着水果和坚果”，跨模态检索系统就会返回对应的图片。引入语音模态后的三模态预训练模型，实现了三模态的统一表示和互相生成，实现了“以图生音”和“以音生图”的功能。
In the case of text searches, the model has achieved a high level of accuracy, and the semantic nature of the results of the entry and retrieval of the returned images is very relevant, entering & ldquo; the blue bowls on the table are filled with fruits and nuts & rdquao; and the cross-moderative search system returns the corresponding picture. The three-module pre-training model that introduces the speech model has been introduced, resulting in a three-module uniform expression and inter-generation, which achieves & & ldquao; & & & & & ldquao; the functions of the audio map & rdquao;
“以图生音”是指输入一个视频或者输入一张图片，系统自动生成语音。“以音生图”可以实现输入语音后，系统自动生成一张图片，例如输入语音：一个人在沙滩上放风筝，系统就会自动生成相对应的图片。
& & ldquo; to type & rdquao; to enter a video or a picture, the system automatically generates a voice. & & & & & rdquao; to enter a voice, the system automatically generates a picture, e.g. to enter a voice: if a person puts a kite on a beach, the system automatically produces a corresponding picture.
随着内容生成、显示和交互技术的发展，数字原生的内容会越来越多，同时数字世界和物理世界的体验也在逐步融合。华为云上数字内容生产线MetaStudio，给各个行业提供云上的媒体基础设施。MetaStudio系统包含核心引擎、平台和流水线三层：引擎层提供华为自主知识产权的云渲染引擎和空间引擎；平台层提供云会议、云桌面、数字资产管理、媒体AI能力，并提供业界首个4K高清云桌面，支持60帧每秒的高帧率，可以替代设计行业目前大量使用的通用工作站；流水线层提供建模、渲染、知识理解、展会、直播等云应用生成能力。这些核心平台和引擎一起支撑多种行业场景中的数字内容的高效生成和运行。MetaStudio已经在数字主持人、数字家装、云上电影制作、虚拟展会等多个场景落地。
With the development of content generation, display and interactive technologies, digital raw content will become more numerous, and the experience of the digital and physical worlds will be gradually integrated. China provides a digital content production line, MetaStudio, which provides the media infrastructure for various industries. The MetaStudio system consists of three layers of core engines, platforms, and waterlines: the engine layer provides a cloud-recording engine and space engine that can serve as autonomous intellectual property rights; the platform layer provides cloud conferences, cloud desktops, digital asset management, media AI capabilities, and the first 4K high cloud desktop in industry, supports a 60-second high frame ratio, which can replace the generic workstations currently used by the design industry; the stream layer provides modelling, dyeing, knowledge understanding, exhibition, live broadcast, etc. The core platforms and engines together support the efficient generation and operation of digital content in a variety of industry landscapes.
随着算力的充沛和模型的成熟，AI开始进入千行百业核心生产系统，产生实际价值。但是，AI在千行百业的落地过程中，都会面临“AI算力稀缺、AI人才短缺、AI开发难、AI行业应用难”这四大难题。
As the arithmetic and model matures, AI begins to move into the 100-way core production system, generating real value. But, in the course of its lifetime, AI faces & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & &.
华为AI开发平台ModelArts持续构建大模型训练、推理优化能力；全面支持MLOps，构建AI开发的持续、协同迭代能力；同时积极打造开放的AI生态，接入生态开发工具，拉通AI落地的供、需、学三方。从算力资源调度、AI业务编排、AI资产管理以及AI应用调试、部署管理，ModelArts开发平台提供数据处理、算法开发、模型训练、模型管理、模型部署等AI模型开发全流程技术能力。
China continuously builds large model training and reasoning optimization capabilities for the AI development platform ModelArts; supports MLOps in a comprehensive manner to build the continuous, synergetic capacity developed by the AI; and is active in building open AI ecology with access to eco-development tools that connects the supply, demand, and learning triads where AI is located. From arithmetic resource movement, AI business organization, AI asset management, and AI application calibration, deployment management, the ModelArts development platform provides full process technical capacity for the development of AI models such as data processing, algorithm development, model training, model management, model deployment, etc.
基于MLOps的理念，华为AI开发平台ModelArts的Workflow工具提供运行记录、监控、持续运行等功能。通过Workflow工具，ModelArts开发平台实现AI开发、运行及运行后监测的全生命周期管理，加快AI开发到落地的迭代速度，以及效果体验的持续优化。同时，开发者可通过华为云AI实训社区AI Gallery实现Workflow这类AI资产的快速构建与能力共享。
Based on the idea of MLOps, China provides running records, monitoring, and running on a continuous basis to the Workflow tool of the AI development platform ModelArts. Through the Workflow tool, ModelArts develops a life-cycle management of AI development, operation, and post-operation monitoring, accelerating the iterative speed of AI development to the site, and continuous optimization of impact experiences. Meanwhile, developers can quickly build and share the capabilities of the AI assets like Workflow through China to train the cloud AI Galileo community.
过去，AI应用开发者平均需要花费半年时间，来适配云边端不同设备的芯片和系统和AI框架，耗时耗力。ModelArts屏蔽了底层软硬件的差异，实现了AI应用一次开发、全场景部署。在某油气田巡检场景中，其通过ModelArts将AI应用一键部署到AR眼镜、手机和机房服务器等多种设备上，让3D重建和定位识别等高性能AI应用能够结合AR眼镜流畅运行，帮助客户解决应用开发中需要兼顾适配不同芯片、系统和推理框架的难题，缩短管线巡检、设备故障识别等AI应用的开发周期。ModelArts能将跨平台开发适配周期缩短80%，将推理性能提升2~10倍。
In the past, it took an average of six months for an AI application developer to adapt the chips and systems and AI frameworks to different equipments at cloud edges to take time and effort. ModelArts shielded the differences in bottom hardware and hardware and achieved a single development and a full-scale deployment of AI applications. In an oil and gas field inspection scene, it deployed an AI application key to a variety of equipment, such as AR glasses, mobile phones and machine room servers, through ModelArts, enabling 3D high-performance AI applications, such as re-establishment and positioning identification, to run smoothly in conjunction with AR glasses flows, helping clients to address the challenge of matching different chip, system and reasoning frameworks in application development, shortening the development cycle of AI applications, such as pipeline inspections, equipment failure recognition, etc. ModelArts can reduce the fit-in cycle of cross-platform development by 80%, and increase reasoning capabilities by 2-10 times.
面向未来的智能世界，计算领域的关键技术升级和演进，朝着六大关键方向进行创新，包括智能认知、内生安全、绿色集约、多样性计算、多维协同、物理层突破。
For a future intelligent world, critical technologies in the field of computing are upgraded and evolved, and innovations are made in six key directions, including intellectual awareness, endogenous security, green convergence, diversity computing, multidimensional synergy, and physical layer breakthroughs.
ICT技术的创新正在从1.0阶段进入2.0阶段。过去，华为无论在无线、光网络、还是智能手机领域，都有大量的工程和技术创新，为客户带来的极大的商业价值以及产生了巨大的社会价值。创新2.0的核心理念是：基于对未来智能社会的假设和愿景，以及世界级难题的基础理论突破和基础技术的发明，为数字世界的发展持续不断地创造新的技术，解决0到1的问题。面向未来，华为将围绕基础科学和前沿技术进行突破性研究，以愿景和假设作为牵引，提出产业需求并攻克世界级的难题。推动产学研创新协同，共创共建，让创新成功持续点亮世界。
Innovation in ICT technology is moving from phase 1.0 to phase 2.0. In the past, China has had a great deal of engineering and technological innovation in the fields of wireless, light networks, and smartphones, bringing great commercial and social value to its clients. The central idea of innovation 2.0 is to create new technologies for the development of the digital world and to solve zero-to-one problems on a continuous basis, based on assumptions and visions for a future intelligent society, as well as on the invention of basic theoretical breakthroughs and basic technologies at the world level.
在通向智能元宇宙世界的道路上， ICT产业下一个十年，联接数量将达到千亿级，宽带速度每人将达到10Gbps，算力和存储能力将提升100倍，可再生能源的使用将超过50%。围绕信息和能量的产生、传送、处理和使用，技术需要不断演进。基于这些预测与假设，华为发布了未来十年的九大技术挑战与研究方向。
On the road to the smart-wing universe, the ICT industry will reach hundreds of billions of levels of connectivity in the next decade, broadband will reach 10 Gbps per person, a 100-fold increase in arithmetic and storage capacity, and more than 50% in renewable energy use. Technology needs to evolve around the generation, transmission, processing and use of information and energy. Based on these projections and assumptions, China has launched nine technological challenges and research directions for the next decade.
① 定义5.5G，支撑未来千亿规模的多样性联接；
1 The definition of 5.5G, which underpins the future multiplicity of hundreds of millions of scales;
② 在纳米尺度上驾驭光，实现光纤容量指数级增长；
To manage light on nanoscales and achieve fibre-optic capacity index-level growth;
③ 走向产业互联，优化网络协议；
3 Towards industrial interconnection and optimization of network protocols;
④ 通用算力远远跟不上智能世界的需求，必须打造超级算力；
4 Universal numeracy is far from keeping up with the needs of the intelligent world and must create super numeracy;
⑤ 从海量多模态的数据中高效地进行知识提取，实现行业AI的关键突破；
5 Efficient knowledge extraction from large-volume multi-module data to achieve a critical breakthrough for industry AI;
⑥ 突破冯诺依曼限制，构建百倍密度增长的新型存储；
6 Breaking the von Neumann restrictions and building new storage with a 100-fold increase in density;
⑦ 将计算与感知结合，实现多模交互的超现实体验；
7 Combining calculations with perception to achieve a hyper-real experience of multi-mode interactions;
⑧ 通过连续性的健康监测实现主动健康管理；
Active health management through continuous health monitoring;
⑨ 构建智慧能源互联网，实现绿色发电、绿色储电和绿色用电。
9 Building the smart energy Internet to achieve green power generation, green storage and green electricity use.
通过关键ICT技术的创新和突破，我们将会实现联接更强、计算更快、能源更绿，最终通向智能元宇宙世界。
Through innovation and breakthroughs in key ICT technologies, we will achieve stronger connectivity, faster computing and greener energy, eventually leading to the smart-wing universe.
注册有任何问题请添加 微信：MVIP619 拉你进入群
打开微信扫一扫
添加客服
进入交流群
发表评论