Intel will provide a broad product portfolio covering data centers, IoT devices and software to promote artificial intelligence to serve enterprises and society
Intel released its artificial intelligence strategy, committed to promoting breakthroughs in artificial intelligence performance bottlenecks, technology popularization, and maximizing social benefits;
Intel announced the launch of the industry's most comprehensive data center computing product portfolio for artificial intelligence: Intel® Nervana™ platform;
Intel aims to shorten the training time of deep learning models to 1/100 of that of GPU solutions in the next three years;
Intel builds an open artificial intelligence ecosystem to promote the realization of greater innovations through a series of developer tools with ease of use and cross-platform compatibility.
Intel Corporation announced the launch of a series of new products, technologies and related investment plans covering from the front end to the data center, aimed at expanding the development space of artificial intelligence (AI) and accelerating its development speed. Intel believes that artificial intelligence will change the way businesses operate and how humans interact with the world. To this end, Intel is integrating its advantages to provide artificial intelligence with the widest range of technology options, and fully release its potential in various fields, including smart factories, drones, sports, fraud detection, and driverless cars.
Intel CEO Ke Zaiqi shared Intel's insights on the prospects and complexity of artificial intelligence. He pointed out that artificial intelligence needs the support of many leading technologies and a larger ecosystem that far exceeds the scope of initial adopters. With the increasing complexity of the current algorithms and the continuous increase of required data sets, Ke Zaiqi said that Intel has a deep insight into these needs and is fully capable of providing the required technology to drive this computing transformation.
"Intel has a unique advantage in enabling and accelerating the innovation of artificial intelligence," said Ke Zaiqi. "Intel has been committed to the development of artificial intelligence, and has invested a lot of technical resources and developers to promote artificial intelligence in enterprises and society. Applications."
Robust Intel artificial intelligence platform
Intel announced that it will launch the industry's comprehensive and leading artificial intelligence product portfolio-Intel ® Nervana™ platform. The product portfolio aims to improve the speed and ease of use of artificial intelligence applications. It is an excellent foundation for building highly optimized artificial intelligence solutions. It can help more data experts to solve the world's extremely difficult problems in technology based on industry standards. challenge.
At present, Intel supports 97% of the world’s data center servers running artificial intelligence workloads, and at the same time provides an extremely flexible and performance-optimized solution product portfolio, including Intel® Xeon® processors, Intel® Xeon Pong Core™ processors, accelerators optimized for workloads such as FPGAs, and technological innovations acquired from Nervana.
Intel also announced more details on how to integrate Nervana's breakthrough technology into its existing product roadmap. Intel will test the first chip (code-named Lake Crest) in the first half of 2017 and sell it to major customers in the second half of the year. In addition, Intel has added a new product (code-named Knights Crest) to the roadmap, which tightly integrates Nervana's innovative technology with the industry-leading Intel Xeon processor. In addition, the Lake Crest processor is specifically optimized for neural networks to provide extremely high performance for deep learning and provide unprecedented computational density through high-speed interconnection networks.
“We look forward to Intel’s Nervana platform that can bring performance advantages that break through the bottleneck, which will greatly reduce the time required to train complex neural networks,” said Ms. Bo Anna, Intel’s executive vice president and general manager of the data center business unit. “Intel is expected to be in 2020 The former will increase performance by 100 times, thus greatly accelerating the pace of innovation in the emerging field of deep learning."
Ms. Bo Anna also announced that compared to the previous generation of processors, the next-generation Intel Xeon Phi processors (code-named Knights Mill) can increase the deep learning performance by 4 times and are scheduled to be launched in 2017. In addition, Intel announced that it has now provided the initial version of the next-generation Intel Xeon processor (code-named Skylake) to specific cloud service provider partners. The processor uses Intel’s advanced vector instruction set AVX-512 integrated acceleration technology, which is extremely Greatly enhance the inference performance of machine learning workloads. The platform series is expected to be officially released in mid-2017, and other enhancements and configurations will be added at that time to fully meet the needs of different customer groups.
Work with Google* to establish a cloud alliance to help popularize artificial intelligence
In addition to chip-level innovation, Intel also disclosed other artificial intelligence layouts, including Intel Saffron Technology™, a leading solution designed to provide customers with business insights. The Saffron platform uses memory-based reasoning technology and a transparent analysis method for heterogeneous data. This technology is very suitable for small devices. It can realize local intelligent analysis in the Internet of Things environment and promote the development of front-end and back-end collaborative artificial intelligence.
To simplify the deployment of artificial intelligence, Intel has also introduced a general intelligence API that supports Intel's distributed processor portfolio from the edge to the cloud platform, as well as Intel ® RealSense™ cameras and Movidius* visual processing units (VPUs).
In addition, Intel and Google officially announced the establishment of a strategic alliance between the two parties to help corporate IT departments build open, flexible, and secure multiple cloud infrastructures based on business needs. The cooperation between the two parties includes technology integration focusing on Kubernetes* (containers), machine learning, security and the Internet of Things.
In order to further promote artificial intelligence research and strategy implementation, Intel also announced the establishment of the Intel Nervana Artificial Intelligence Committee, which is mainly composed of leading thought leaders in the industry and academia. Intel announced that the four founding members are Yoshua Bengio (Yoshua Bengio) from the University of Montreal, Bruno Olshausen (Bruno Olshausen) and Rabeai (Jan Rabaey) and Ron Dror from Stanford University.
Intel will also spare no effort to promote the real landing of artificial intelligence. To this end, Intel has launched the Intel Nervana Artificial Intelligence Academy to developers to facilitate their acceptance of relevant training and tools. Intel also introduced the Intel Nervana graphics compiler to accelerate the development of deep learning frameworks in Intel chips.
In addition to the Academy of Artificial Intelligence, Intel also announced a partnership with Coursera*, the world's leading educational institution, to provide a series of artificial intelligence online courses for the academic community. Intel will also co-host a Kaggle competition with Mobile ODT* (expected to be launched in January 2017). The academic community can take this opportunity to use its artificial intelligence skills to solve practical socio-economic problems, such as using artificial intelligence to photograph soft tissues in developing countries Imaging photos to identify early cervical cancer.
"Intel can provide major technologies to promote the artificial intelligence revolution, and ultimately, the entire industry and society must work closely to explore the greatest potential of artificial intelligence." said Fei Daoming, Intel's senior vice president and general manager of the Software and Services Division.
By continuously launching innovative products and support projects covering front-end and data centers, Intel has sufficient technical strength and ecosystem support to promote the innovation and popularization of artificial intelligence.
Artificial intelligence increases social well-being
Recently, Intel demonstrated some of its investment and cooperation actions to help increase the positive impact of artificial intelligence on the world, including:
Intel invested 25 million US dollars in the American Broad Institute to promote high-performance genetic analysis and computing research. In the five-year collaboration, researchers and software engineers at the Intel-Broad Genetic Data Engineering Center will build, optimize, and widely share new tools and infrastructure to help scientists integrate and process genetic data. The goal of the project is to optimize best practices in genetic analysis hardware and software, so that the public can access and use research data located in private, public, and hybrid cloud platforms.
Intel is the founding partner of Hack Harassment*. The goal of this cooperative action is to reduce the spread and impact of online harassment. The action will evaluate the applicability of artificial intelligence technology tools and work hard to develop intelligent algorithms to detect and stop network harassment. In the future, this feature will be released as an open source API, benefiting a wider range of applications.
Intel is also a major partner of the National Center for Missing and Exploited Children* (NCMEC). NCMEC is a non-profit organization that aims to help find missing children, reduce child abuse, and prevent child abuse. Intel will provide artificial intelligence technology and provide consulting suggestions to help analysts speed up the completion of key tasks and respond to reported child sexual assaults in a timely manner.
