Embedded Computer Systems: COM and AI Edge Computing Technology

Accelerated Generative AI Development with NVIDIA Jetson AGX Orin-Based Edge Computer

Advantech's MIC-733-AO now runs a range of advanced AI capabilities including large language models, vision language models, and vision transformers By Mike Ball / 13 Jun 2024
Follow UST

Advantech has confirmed that it has successfully demonstrated generative AI using the company’s MIC-733-AO edge computing system, based on the NVIDIA Jetson AGX Orin platform.

This demonstration shows that AI developers can benefit from greater immediacy, enhanced security, and reduced internet costs compared to cloud-based generative AI. The generative AI models from the Jetson AI lab offers a free foundational AI model, enabling customization and significantly reducing development time with the MIC-733-AO Edge Generative AI System.

Advantech’s MIC-733-AO Edge Generative AI System now efficiently runs large language models (LLMs), vision language model (VLM), vision transformers (ViT), and image generation and Llamaspeak at the edge, thanks to the NVIDIA Jetson AGX Orin. The platform’s unique offerings span multiple layers, including accelerated computing, vital AI software, and pre-trained models, facilitating the creation, customization, and deployment of generative AI models suitable for various applications and locations.

With Generative AI models from the Jetson AI lab, developers can leverage the retrieval-augmented generation (RAG) technique, connecting LLMs and VLMs to a company’s knowledge base. This approach eliminates the need for traditional AI model training, allowing simple conversational interactions with the system to yield AI inferences, significantly shortening development time.

The Advantech MIC-733-AO features high AI performance and multiple video interfaces, such as IP cameras, USB/gigE industrial cameras, GMSL cameras, LiDAR, and more, enabling edge-based Generative AI model operation. Its exceptional thermal design prevents hot crashes, even when the NVIDIA Jetson AGX Orin operates at full capacity.

Generative AI is revolutionizing human-computer interaction by creating new content from various inputs and outputs, such as text and images. The shift in demand from the cloud to the edge positions GenAI as ideal for high-responsiveness and low-latency environments.

Edge-based Generative AI offers several advantages over cloud-based solutions:

Immediate Response – Edge computing systems, integrating data sources like cameras and sensors, enable prompt GenAI responses without the delays associated with cloud uploads.
Information Security – Data remains within the edge system, avoiding the cybersecurity risks associated with cloud data transmission.
Stable Network Connection – Operable without a wireless network, the edge system stores data locally until network restoration, unlike cloud-based AI, which ceases function without a network.
Low Transmission Costs – Edge AI incurs no transmission costs if backend communication is unnecessary, needing only low-bandwidth transmission, more cost-effective than the higher bandwidth costs of cloud uploads.

Advantech will be performing a Generative AI demo at the 2024 NVIDIA GTC AI conference. Interested parties can also access an application note used as the basis for the demonstration here.

To learn more, contact Advantech: Visit Website Send Message View Supplier Profile
Posted by Mike Ball Mike Ball is our resident technical editor here at Unmanned Systems Technology. Combining his passion for teaching, advanced engineering and all things unmanned, Mike keeps a watchful eye over everything related to the unmanned technical sector. With over 10 years’ experience in the unmanned field and a degree in engineering, Mike’s been heading up our technical team here for the last 8 years. Connect & Contact