Issue link: https://resources.mouser.com/i/1437726
34 REIMAGINING WHAT'S NEXT Additional Intel AI Resources The oneAPI tools mentioned thus far are the core of Intel's offering for developing AI-centric applications. A few other tools to be aware of include: 1. AI DevCloud provides a cloud-based sandbox for data center, FPGA, and edge device workloads so the developer can try their applications using the AI Analytics Toolkit (via the Intel oneAPI programming model) on virtualized Intel hardware before making any significant investments. Intel oneAPI allows any application, including AI, targeted to any 'XPU,' be it a CPU, GPU, FPGA, or any custom accelerator chip. Figure 7: AI DevCloud lets developers try out Intel hardware and software virtually before making investment and purchasing decisions. (Source: Intel) 2. Model Zoo for Intel Architecture gives developers access to numerous pre-built ML models for quick turnaround application development. 3. Intel® Quartus® Prime Design software can be usedt o target ML/DL models to Intel FPGAs when used in conjunction with OpenVINO. 4. Intel AI Builders is a community of independent software vendors (ISVs), system integrators (SIs), original equipment manufacturers (OEMs), and enterprise end-users who are leading the adoption of AI technologies within the Intel ecosystem. Ultimately AI enables automation that is both adaptable and self-improving, much like our biological brains. For Intel and their oneAPI initiative, the advantage of providing ML framework developers with free and open-source optimized mathematical and analytical libraries is that it encourages AI application developers and cloud service providers who adopt the optimized frameworks significant incentive to select Intel hardware. By adopting a stack of fully optimized tools (from the hardware all the way to high-level software development kits), developers can reduce the time to market while also achieving unparalleled performance. Furthermore, the open-source licensing and relatively low upfront costs allow a wide range of individuals and organizations to take advantage of these ML/DL technologies. Thus the possibility of exciting new AI-powered innovations can come from anywhere. Figure 6: OpenVINO helps to easily integrate NN models on low-cost embedded hardware such as Intel's Movidius Neural Computer Stick. (Source: Intel)