Lenovo has developed a framework for deploying Generative AI in high demand urban environments. The Lenovo Hybrid AI Advantage with NVIDIA framework includes new patent-pending technologies visual insights network for AI (Lenovo VINA) and edge guarding.

Both Lenovo edge guarding and Lenovo VINA leverage NVIDIA Metropolis and NVIDIA AI Blueprints as reference workflows for generative AI use cases. NVIDIA AI Blueprint for video search and summarisation enables the deployment of AI agents trained to analyse visual data, summarise massive data repositories, and produce answers to user prompts. According to Lenovo, the visual agents can understand large repositories of live or archived video to perceive and reason against action occurring across any space with sensor infrastructure and accelerate incident search. The company sees smart city applications in traffic management, parking, curbside management, and public safety.

Using the NVIDIA AI Blueprint for video search and summarisation, the visual language model (VLM) technology alongside VINA can analyse large volumes of visual data and prioritise high-value information for quick retrieval, reducing storage needs and operational costs.

Lenovo’s edge guarding will enable an autonomous, resilient local network among edge devices found across city infrastructure, including traffic monitors, sensors, and city-wide cameras that can operate independently of cloud connections during disruptions. Edge guarding’s GPU allocation model prioritises tasks according to urgency, allowing high-stakes operations to continue without interruptions.

Lenovo’s edge guarding predictive security model, powered by NVIDIA’s machine learning capabilities from NVIDIA AI Blueprints, is designed to strengthen the city’s defences against cyber threats by dynamically adjusting task priorities based on potential vulnerabilities. If a network outage occurs, edge guarding will establish a local network, keeping critical infrastructure, such as public safety systems, operational.