새로운 프로그램을 위한 혁신적이며, 협업이 가능한 동기화된 프로그램 관리
In recent years, advances in hardware technology and software techniques have made it possible and practical for machine learning to provide better solutions to many real-world problems, even on embedded devices.
Nevertheless, embedded machine learning presents a unique set of challenges, including hardware resource requirements and performance considerations, the need for integration of specialized software technologies, and not least the learning curve – including a dependency on skill sets that are often not present in existing development teams.
Siemens Embedded's platform solutions and professional services enable our customers to get started quickly and easily with machine learning and allow them to focus on how machine learning can add value to their products, without first having to solve a bewildering and time-consuming array of complex platform-level challenges.
In recent years, advances in hardware technology and software techniques have made it possible and practical for machine learning to provide better solutions to many real-world problems, even on embedded devices.
Nevertheless, embedded machine learning presents a unique set of challenges, including hardware resource requirements and performance considerations, the need for integration of specialized software technologies, and not least the learning curve – including a dependency on skill sets that are often not present in existing development teams.
Siemens Embedded's platform solutions and professional services enable our customers to get started quickly and easily with machine learning and allow them to focus on how machine learning can add value to their products, without first having to solve a bewildering and time-consuming array of complex platform-level challenges.
Machine learning solutions are typically built upon open-source framework software. The Siemens Embedded Sokol™ Machine Learning Add-on solution includes an integration of the most popular machine learning engine for production embedded projects, TensorFlow Lite.
Other engines and frameworks can be provided on request. The framework is integrated with the Siemens Embedded Linux platforms Sokol™ Omni OS and Sokol™ Flex OS, optimized for the hardware target, rigorously tested, commercially supported, and monitored for CVEs alongside the rest of our platform software.
Customers can be assured that the machine learning platform software upon which their applications are built is always robust, reliable and secure, from right out of the box.
For customers that need additional help, our team of machine learning integration specialists offer both professional services and long-term premium support. Examples of work undertaken include integration of sensors and processing pipelines to drive machine learning inference, integration of alternative machine learning frameworks, integration of artificial intelligence accelerators or other hardware devices, and performance optimization and tuning.
Sokol™ Machine Learning Add-on also benefits from integration with our other tools and frameworks.
Customers can easily build and debug platforms and applications from within the Sourcery™ Codebench™ IDE, and visualize and analyze performance of machine learning solutions using Sourcery™ Analyzer, allowing complex system-level performance issues to be diagnosed.
Machine learning solutions can be monitored and managed from the cloud using Sokol™ IoT Framework, including data acquisition, analysis of system performance, and update of machine learning models in the field.
Software and expertise enabling customers to get started quickly and easily with machine learning, without first having to solve a bewildering and time-consuming array of complex platform-level challenges. Along with key open-source technologies and productivity tools, security vulnerability monitoring and support, customers can move forward into machine learning with confidence.