微信号

功能介绍

Artificial Intelligence

人工智能

Artificial Intelligence (AI) has long been a strategic priority for NIST, also representing a toolkit to remarkably enhance productivity across all areas of our research and development, as well as toward advanced manufacturing applications. An important goal for NIST is to develop the foundation for confidence and trust in AI that results in new research outcomes and an expanded commercial marketplace. 

人工智能(AI)长期以来一直是NIST的战略重点，也代表了一个工具包，可以显着提高我们所有研究和开发领域的生产力，以及先进的制造应用。NIST的一个重要目标是建立对人工智能的信心和信任基础，从而产生新的研究成果和扩大的商业市场。

International investment in AI is also exploding, and companies, governments and policy makers around the globe are seeking answers that can provide greater confidence in AI technologies. NIST’s study and deployment of AI methods, tools, and standards can provide the basis for confidence and trust that is essential for adoption of these technologies.

国际上对人工智能的投资也在爆炸式增长，全球各地的公司、政府和政策制定者都在寻求能让人们对人工智能技术更有信心的答案。NIST对人工智能方法、工具和标准的研究和部署可以为采用这些技术提供信心和信任的基础。

NIST has made significant contributions to the fields of machine learning (ML) and AI over the years. For example, the MNIST database, a dataset of handwritten digits, is among the most widely used standardized datasets in the U.S. and around the world for training and testing AI systems. 

多年来，NIST在机器学习(ML)和人工智能领域做出了重大贡献。例如，MNIST数据库是一个手写数字数据集，是美国和世界各地用于训练和测试人工智能系统的最广泛使用的标准化数据集之一。

NIST scientists worked with the Defense Advanced Research Projects Agency to develop and deploy smartphone-based systems that enabled U.S. marines to seamlessly converse with native Pashto speaking Afghans. These technology developments have also facilitated rapid commercialization of phone-based voice translation systems such as Microsoft Bing and Google Translate.

NIST的科学家们与国防高级研究计划局合作，开发并部署了基于智能手机的系统，使美国海军陆战队能够与当地讲普什图语的阿富汗人无缝交谈。这些技术的发展也促进了基于电话的语音翻译系统的快速商业化，如微软必应和谷歌翻译。

Today, NIST’s efforts in AI are focused along three primary areas of effort:

如今，NIST在人工智能方面的努力主要集中在三个主要领域:

First, NIST is addressing fundamental questions about the use of AI. NIST has launched an effort to convene the community around key concepts of trustworthy AI, seeking to develop ways to measure, define, and characterize concepts around the accuracy, reliability, privacy, robustness, and explainability of AI systems. Some examples of NIST work in this space include:

首先，NIST正在解决有关人工智能使用的基本问题。NIST发起了一项努力，围绕可信赖的人工智能的关键概念召集社区，寻求开发方法来衡量、定义和表征人工智能系统的准确性、可靠性、隐私性、鲁棒性和可解释性等概念。NIST在这一领域的一些工作包括:

• In November, the NIST National Cybersecurity Center of Excellence (NCCoE) issued a draft NIST Internal Report, “A Taxonomy and Terminology of Adversarial Machine Learning.”

• 去年11月，NIST国家网络安全卓越中心(NCCoE)发布了一份NIST内部报告草案，“对抗性机器学习的分类和术语”。
  

• In December, NIST issued a report on the performance of face recognition software tools in identifying people of varied sex, age and racial backgrounds: “Face Recognition Vendor Test (FRVT) Part 3: Demographic Effects” (NISTIR 8280). Such data is intended to provide valuable insights to policymakers, developers and end users about the limitations and appropriate uses of currently available AI tools.

• 去年12月，NIST发布了一份关于人脸识别软件工具在识别不同性别、年龄和种族背景的人方面的表现的报告:“人脸识别供应商测试(FRVT)第3部分:人口统计效应”(NISTIR 8280)。这些数据旨在为政策制定者、开发人员和最终用户提供有关当前可用人工智能工具的局限性和适当使用的宝贵见解。
  

• NIST and its NCCoE are planning to launch a testbed to evaluate AI vulnerabilities.

• NIST及其NCCoE计划推出一个测试平台来评估人工智能漏洞。
  

• Later in January, NIST intends to release a set of draft “Principles of Explainable AI” for public comment.

• 1月晚些时候，NIST打算发布一套“可解释人工智能原则”草案，征求公众意见。
  

• NIST is organizing a workshop to convene stakeholders to explore issues of bias in machine-learning based face and speech recognition algorithms.

• NIST正在组织一个研讨会，召集利益相关者探讨基于机器学习的人脸和语音识别算法中的偏见问题。

Secondly, NIST is heavily engaged in using AI across its research portfolio in a host of areas including biometrics, advanced materials discovery, smart manufacturing systems, and the design and characterization of engineered biological systems as just a few examples. 

其次，NIST在许多领域的研究组合中大量使用人工智能，包括生物识别、先进材料发现、智能制造系统以及工程生物系统的设计和表征，这只是几个例子。

Additionally, the outputs of NIST research in general, especially in the terms of well-characterized data sets, as well as our work in advanced microelectronic systems, will help advance the field of AI. These tools will enable researchers to better train and understand AI systems, including the design and manufacture of next-generation hardware required to reliably and safely run AI systems. Some recent examples of NIST effort in this space include:

此外，NIST研究的总体成果，特别是在特征良好的数据集方面，以及我们在先进微电子系统方面的工作，将有助于推动人工智能领域的发展。这些工具将使研究人员能够更好地训练和理解人工智能系统，包括设计和制造可靠、安全运行人工智能系统所需的下一代硬件。NIST最近在这一领域的一些努力包括:

• NIST researchers are working on ways to utilize AI to automate vulnerability assessments for digital infrastructure and to produce vulnerability ratings using the industry-standard Common Vulnerability Scoring System.

• NIST的研究人员正在研究如何利用人工智能自动对数字基础设施进行漏洞评估，并使用行业标准的通用漏洞评分系统生成漏洞评级。
  

• In advanced materials discovery, NIST has created a high-fidelity database, Joint Automated Repository for Various Integrated Simulations, density functional theory (JARVIS-DFT), with more than 30,000 materials and 500,000 properties to be used as training data that will help accelerate the development of new materials.

• 在先进材料发现方面，NIST创建了一个高保真数据库，即各种集成模拟、密度泛函理论联合自动化存储库(JARVIS-DFT)，其中包含超过30,000种材料和500,000种特性，将用作训练数据，有助于加速新材料的开发。
  

• In wireless spectrum analysis, NIST is creating a curated radio frequency (RF) signal database to aid in the development of machine learning models for signal detection and classification. These datasets, which include radar signals similar to those planned for the 3.5 GHz band and include noise and interference, can be used to train and evaluate AI detectors to enable federal-commercial spectrum sharing.

• 在无线频谱分析方面，NIST正在创建一个精心策划的射频(RF)信号数据库，以帮助开发用于信号检测和分类的机器学习模型。这些数据集包括类似于3.5 GHz频段的雷达信号，包括噪声和干扰，可用于训练和评估人工智能探测器，以实现联邦和商业频谱共享。
  

• In manufacturing, NIST is applying AI in its study of agility performance of robotic systems in manufacturing environments so that robots can “learn” behaviors to operate effectively in today’s factories. Later this month, NIST will launch our fourth annual Agile Robotics for Industrial Automation Competition, offering cash prizes to the teams whose robots perform the best in a simulated environment.

• 在制造业方面，NIST正在将人工智能应用于机器人系统在制造环境中的敏捷性能研究，以便机器人能够“学习”行为，在当今的工厂中有效地运作。本月晚些时候，NIST将启动第四届年度工业自动化敏捷机器人竞赛，为机器人在模拟环境中表现最佳的团队提供现金奖励。

Finally, standards engagement is a key element of NIST’s mission, and we are deeply involved in multiple standards development bodies around the world. We are working with industry, government, and academia to establish governing principles and develop standards and identify best practices for the design, construction, and use of AI systems. It is vitally important for the U.S. to have a strong, persuasive, and consistent voice with the relevant standards organizations around the world.

最后，标准参与是NIST使命的一个关键要素，我们深入参与了世界各地的多个标准开发机构。我们正在与工业界、政府和学术界合作，建立治理原则，制定标准，并确定人工智能系统的设计、建设和使用的最佳实践。对美国来说，在世界各地的相关标准组织中拥有一个强有力的、有说服力的、一致的声音是至关重要的。

• In August 2019, NIST released the report “U.S. Leadership in AI: A Plan for Federal Engagement in Developing Technical Standards and Related Tools” in response to the Executive Order (EO) 13859 directing NIST to issue a plan for federal engagement in the development of technical standards and related tools in support of reliable, robust, and trustworthy systems that use AI technologies. The plan identifies nine areas of focus for AI standards and urges that the federal government commit to deeper, consistent, long-term engagement in AI standards development.

• 2019年8月，美国国家标准与技术研究所(NIST)发布了题为《美国《人工智能领导力:联邦参与开发技术标准和相关工具的计划》是为了响应第13859号行政命令，该命令指示NIST发布联邦参与开发技术标准和相关工具的计划，以支持使用人工智能技术的可靠、稳健和可信赖的系统。该计划确定了人工智能标准的九个重点领域，并敦促联邦政府致力于更深入、一致、长期地参与人工智能标准的制定。
  

• Twelve NIST experts are currently involved in the joint International Standards Organization (ISO) / International Electrotechnical Committee (IEC) Joint Technical Committee JTC 1, Subcommittee (SC) 42 on Artificial Intelligence, and NIST is the convener for the Big Data work effort in SC 42. NIST works with many companies (including Google, Intel, Microsoft, and Oracle), other federal agencies, and academia to develop U.S. consensus positions on the U.S. Technical Advisory Group for SC 42, supported by the International Committee for Information Technology Standards.

• 目前有12名NIST专家参与了国际标准组织(ISO) /国际电工委员会(IEC)联合技术委员会JTC 1，人工智能分委员会(SC) 42, NIST是SC 42中大数据工作的召集人。NIST与许多公司(包括Google、Intel、Microsoft和Oracle)、其他联邦机构和学术界合作，在国际信息技术标准委员会的支持下，就SC 42的美国技术咨询小组制定美国的共识立场。
  

• NIST staff are participating in over a dozen other AI standards activities in various standards development organizations, including the American Society for Mechanical Engineers (ASME), the Institute of Electrical and Electronics Engineers (IEEE), and ISO/IEC. These activities cover topics such as computational modeling for advanced manufacturing, ontologies for robotics and automation, personal data privacy, and algorithmic bias.

• NIST的工作人员正在各种标准开发组织中参与十多个其他人工智能标准活动，包括美国机械工程师协会(ASME)、电气和电子工程师协会(IEEE)和ISO/IEC。这些活动涵盖了先进制造业的计算建模、机器人和自动化的本体、个人数据隐私和算法偏见等主题。
  

NIST’s capabilities, ranging from fundamental research to the delivery of the technical foundations of emerging technologies, make it a valuable asset in establishing and maintaining U.S. leadership in AI technologies.

NIST的能力，从基础研究到新兴技术的技术基础交付，使其成为建立和保持美国在人工智能技术领域领导地位的宝贵资产。

扫一扫

微信号

功能介绍

Artificial Intelligence

人工智能

Artificial Intelligence (AI) has long been a strategic priority for NIST, also representing a toolkit to remarkably enhance productivity across all areas of our research and development, as well as toward advanced manufacturing applications. An important goal for NIST is to develop the foundation for confidence and trust in AI that results in new research outcomes and an expanded commercial marketplace. 

人工智能(AI)长期以来一直是NIST的战略重点，也代表了一个工具包，可以显着提高我们所有研究和开发领域的生产力，以及先进的制造应用。NIST的一个重要目标是建立对人工智能的信心和信任基础，从而产生新的研究成果和扩大的商业市场。

International investment in AI is also exploding, and companies, governments and policy makers around the globe are seeking answers that can provide greater confidence in AI technologies. NIST’s study and deployment of AI methods, tools, and standards can provide the basis for confidence and trust that is essential for adoption of these technologies.

国际上对人工智能的投资也在爆炸式增长，全球各地的公司、政府和政策制定者都在寻求能让人们对人工智能技术更有信心的答案。NIST对人工智能方法、工具和标准的研究和部署可以为采用这些技术提供信心和信任的基础。

NIST has made significant contributions to the fields of machine learning (ML) and AI over the years. For example, the MNIST database, a dataset of handwritten digits, is among the most widely used standardized datasets in the U.S. and around the world for training and testing AI systems. 

多年来，NIST在机器学习(ML)和人工智能领域做出了重大贡献。例如，MNIST数据库是一个手写数字数据集，是美国和世界各地用于训练和测试人工智能系统的最广泛使用的标准化数据集之一。

NIST scientists worked with the Defense Advanced Research Projects Agency to develop and deploy smartphone-based systems that enabled U.S. marines to seamlessly converse with native Pashto speaking Afghans. These technology developments have also facilitated rapid commercialization of phone-based voice translation systems such as Microsoft Bing and Google Translate.

NIST的科学家们与国防高级研究计划局合作，开发并部署了基于智能手机的系统，使美国海军陆战队能够与当地讲普什图语的阿富汗人无缝交谈。这些技术的发展也促进了基于电话的语音翻译系统的快速商业化，如微软必应和谷歌翻译。

Today, NIST’s efforts in AI are focused along three primary areas of effort:

如今，NIST在人工智能方面的努力主要集中在三个主要领域:

First, NIST is addressing fundamental questions about the use of AI. NIST has launched an effort to convene the community around key concepts of trustworthy AI, seeking to develop ways to measure, define, and characterize concepts around the accuracy, reliability, privacy, robustness, and explainability of AI systems. Some examples of NIST work in this space include:

首先，NIST正在解决有关人工智能使用的基本问题。NIST发起了一项努力，围绕可信赖的人工智能的关键概念召集社区，寻求开发方法来衡量、定义和表征人工智能系统的准确性、可靠性、隐私性、鲁棒性和可解释性等概念。NIST在这一领域的一些工作包括:

• In November, the NIST National Cybersecurity Center of Excellence (NCCoE) issued a draft NIST Internal Report, “A Taxonomy and Terminology of Adversarial Machine Learning.”

• 去年11月，NIST国家网络安全卓越中心(NCCoE)发布了一份NIST内部报告草案，“对抗性机器学习的分类和术语”。
  

• In December, NIST issued a report on the performance of face recognition software tools in identifying people of varied sex, age and racial backgrounds: “Face Recognition Vendor Test (FRVT) Part 3: Demographic Effects” (NISTIR 8280). Such data is intended to provide valuable insights to policymakers, developers and end users about the limitations and appropriate uses of currently available AI tools.

• 去年12月，NIST发布了一份关于人脸识别软件工具在识别不同性别、年龄和种族背景的人方面的表现的报告:“人脸识别供应商测试(FRVT)第3部分:人口统计效应”(NISTIR 8280)。这些数据旨在为政策制定者、开发人员和最终用户提供有关当前可用人工智能工具的局限性和适当使用的宝贵见解。
  

• NIST and its NCCoE are planning to launch a testbed to evaluate AI vulnerabilities.

• NIST及其NCCoE计划推出一个测试平台来评估人工智能漏洞。
  

• Later in January, NIST intends to release a set of draft “Principles of Explainable AI” for public comment.

• 1月晚些时候，NIST打算发布一套“可解释人工智能原则”草案，征求公众意见。
  

• NIST is organizing a workshop to convene stakeholders to explore issues of bias in machine-learning based face and speech recognition algorithms.

• NIST正在组织一个研讨会，召集利益相关者探讨基于机器学习的人脸和语音识别算法中的偏见问题。

Secondly, NIST is heavily engaged in using AI across its research portfolio in a host of areas including biometrics, advanced materials discovery, smart manufacturing systems, and the design and characterization of engineered biological systems as just a few examples. 

其次，NIST在许多领域的研究组合中大量使用人工智能，包括生物识别、先进材料发现、智能制造系统以及工程生物系统的设计和表征，这只是几个例子。

Additionally, the outputs of NIST research in general, especially in the terms of well-characterized data sets, as well as our work in advanced microelectronic systems, will help advance the field of AI. These tools will enable researchers to better train and understand AI systems, including the design and manufacture of next-generation hardware required to reliably and safely run AI systems. Some recent examples of NIST effort in this space include:

此外，NIST研究的总体成果，特别是在特征良好的数据集方面，以及我们在先进微电子系统方面的工作，将有助于推动人工智能领域的发展。这些工具将使研究人员能够更好地训练和理解人工智能系统，包括设计和制造可靠、安全运行人工智能系统所需的下一代硬件。NIST最近在这一领域的一些努力包括:

• NIST researchers are working on ways to utilize AI to automate vulnerability assessments for digital infrastructure and to produce vulnerability ratings using the industry-standard Common Vulnerability Scoring System.

• NIST的研究人员正在研究如何利用人工智能自动对数字基础设施进行漏洞评估，并使用行业标准的通用漏洞评分系统生成漏洞评级。
  

• In advanced materials discovery, NIST has created a high-fidelity database, Joint Automated Repository for Various Integrated Simulations, density functional theory (JARVIS-DFT), with more than 30,000 materials and 500,000 properties to be used as training data that will help accelerate the development of new materials.

• 在先进材料发现方面，NIST创建了一个高保真数据库，即各种集成模拟、密度泛函理论联合自动化存储库(JARVIS-DFT)，其中包含超过30,000种材料和500,000种特性，将用作训练数据，有助于加速新材料的开发。
  

• In wireless spectrum analysis, NIST is creating a curated radio frequency (RF) signal database to aid in the development of machine learning models for signal detection and classification. These datasets, which include radar signals similar to those planned for the 3.5 GHz band and include noise and interference, can be used to train and evaluate AI detectors to enable federal-commercial spectrum sharing.

• 在无线频谱分析方面，NIST正在创建一个精心策划的射频(RF)信号数据库，以帮助开发用于信号检测和分类的机器学习模型。这些数据集包括类似于3.5 GHz频段的雷达信号，包括噪声和干扰，可用于训练和评估人工智能探测器，以实现联邦和商业频谱共享。
  

• In manufacturing, NIST is applying AI in its study of agility performance of robotic systems in manufacturing environments so that robots can “learn” behaviors to operate effectively in today’s factories. Later this month, NIST will launch our fourth annual Agile Robotics for Industrial Automation Competition, offering cash prizes to the teams whose robots perform the best in a simulated environment.

• 在制造业方面，NIST正在将人工智能应用于机器人系统在制造环境中的敏捷性能研究，以便机器人能够“学习”行为，在当今的工厂中有效地运作。本月晚些时候，NIST将启动第四届年度工业自动化敏捷机器人竞赛，为机器人在模拟环境中表现最佳的团队提供现金奖励。

Finally, standards engagement is a key element of NIST’s mission, and we are deeply involved in multiple standards development bodies around the world. We are working with industry, government, and academia to establish governing principles and develop standards and identify best practices for the design, construction, and use of AI systems. It is vitally important for the U.S. to have a strong, persuasive, and consistent voice with the relevant standards organizations around the world.

最后，标准参与是NIST使命的一个关键要素，我们深入参与了世界各地的多个标准开发机构。我们正在与工业界、政府和学术界合作，建立治理原则，制定标准，并确定人工智能系统的设计、建设和使用的最佳实践。对美国来说，在世界各地的相关标准组织中拥有一个强有力的、有说服力的、一致的声音是至关重要的。

• In August 2019, NIST released the report “U.S. Leadership in AI: A Plan for Federal Engagement in Developing Technical Standards and Related Tools” in response to the Executive Order (EO) 13859 directing NIST to issue a plan for federal engagement in the development of technical standards and related tools in support of reliable, robust, and trustworthy systems that use AI technologies. The plan identifies nine areas of focus for AI standards and urges that the federal government commit to deeper, consistent, long-term engagement in AI standards development.

• 2019年8月，美国国家标准与技术研究所(NIST)发布了题为《美国《人工智能领导力:联邦参与开发技术标准和相关工具的计划》是为了响应第13859号行政命令，该命令指示NIST发布联邦参与开发技术标准和相关工具的计划，以支持使用人工智能技术的可靠、稳健和可信赖的系统。该计划确定了人工智能标准的九个重点领域，并敦促联邦政府致力于更深入、一致、长期地参与人工智能标准的制定。
  

• Twelve NIST experts are currently involved in the joint International Standards Organization (ISO) / International Electrotechnical Committee (IEC) Joint Technical Committee JTC 1, Subcommittee (SC) 42 on Artificial Intelligence, and NIST is the convener for the Big Data work effort in SC 42. NIST works with many companies (including Google, Intel, Microsoft, and Oracle), other federal agencies, and academia to develop U.S. consensus positions on the U.S. Technical Advisory Group for SC 42, supported by the International Committee for Information Technology Standards.

• 目前有12名NIST专家参与了国际标准组织(ISO) /国际电工委员会(IEC)联合技术委员会JTC 1，人工智能分委员会(SC) 42, NIST是SC 42中大数据工作的召集人。NIST与许多公司(包括Google、Intel、Microsoft和Oracle)、其他联邦机构和学术界合作，在国际信息技术标准委员会的支持下，就SC 42的美国技术咨询小组制定美国的共识立场。
  

• NIST staff are participating in over a dozen other AI standards activities in various standards development organizations, including the American Society for Mechanical Engineers (ASME), the Institute of Electrical and Electronics Engineers (IEEE), and ISO/IEC. These activities cover topics such as computational modeling for advanced manufacturing, ontologies for robotics and automation, personal data privacy, and algorithmic bias.

• NIST的工作人员正在各种标准开发组织中参与十多个其他人工智能标准活动，包括美国机械工程师协会(ASME)、电气和电子工程师协会(IEEE)和ISO/IEC。这些活动涵盖了先进制造业的计算建模、机器人和自动化的本体、个人数据隐私和算法偏见等主题。
  

NIST’s capabilities, ranging from fundamental research to the delivery of the technical foundations of emerging technologies, make it a valuable asset in establishing and maintaining U.S. leadership in AI technologies.

NIST的能力，从基础研究到新兴技术的技术基础交付，使其成为建立和保持美国在人工智能技术领域领导地位的宝贵资产。

扫一扫