Baidu Aims Edge Platform at AI Developers
Platform vendors are increasingly focusing on computing at the network edge as those devices sweep up more unstructured data. The goal is to move computing resources closer to data sources to reduce latency and bandwidth usage.
The latest example comes from Chinese AI leader Baidu, which this week rolled out an open-source computing platform aimed at developers seeking to build edge applications. Baudi said those applications could include new features such as AI inference, cloud synchronization, data collection and message distribution, among others.
Beijing-based Baidu (NASDAQ: BIDU) unveiled its OpenEdge platform Wednesday (Jan.9) at this week’s Consumer Electronics Show in Las Vegas.
Described as a “local package component” of Baidu’s Intelligent Edge (BIE) commercial software, OpenEdge looks to capitalize on the growth in Internet of Things devices and growing AI adoption that are fueling demand of edge computing. BIE supports models trained with AI frameworks such as Baidu’s PaddlePaddle open source deep learning and TensorFlow.
Baidu said the package would allow developers to train AI models in the cloud, then deploy them to local devices.
In conjunction with the edge computing rollout, Baidu also announced additional collaboration with Intel Corp. (NASDAQ: INTC) on “AI Box” hardware that incorporates Baidu’s edge computing technologies into vehicle cameras handling video analysis. The partners said the hardware would provision AI applications ranging from road recognition to driver behavior monitoring.
The hardware extends collaboration between Baidu and Intel announced last summer that includes AI projects ranging from FPGA-backed workload acceleration, a deep learning framework based on Xeon scalable processors and implementation of a vision processing unit.
Separately, Baidu and NXP Semiconductors (NASDAQ: NXPI) jointly launched BEI-based “AI Board” with software that could be embedded into cameras, drones and other detection devices. NXP is a key supplier to auto makers, particularly chips used to process sensor data used to guide self-driving cars.
By “providing an open source platform, we have also greatly simplified the process for developers to create their own edge computing applications,” said Watson Yin, general manager of Baidu Cloud.
The edge computing platform is among a growing list of AI development tools released by the Chinese e-commerce giant over the last year. Baidu unveiled a deep learning platform in September dubbed EZDL. The service is being promoted as a way to build custom machine learning models using a drag-and-drop interface. The platform focuses on three primary capabilities: image and sound classification, the latter including voice recognition, along with object detection.