How Edge AI is a Roadmap to Future AI and IoT Trends?

by June 29, 2020

Edge - AI

Why AI Applications in Edge Devices and Computing is the future?

Change has always been integral to development. With fast-evolving technologies, companies, too, need themselves to embrace these for maximized benefits. Artificial Intelligence (AI) moving to edge IoT devices and networks, just like we witnessed computing switch from mainframes to the cloud. And as data continues to grow, we need to opt for data storage and data computation to be located on the device. Companies like Qualcomm, NVIDIA, and Intel are helping us achieve this reality.

While edge site computing systems are much smaller than those found in central data centers, they have matured, and now successfully run many workloads due to immense growth in the processing power of today’s x86 commodity servers. Plus, edge is a better option if an application is latency-sensitive. Better privacy, Security, low latency, and bandwidth are some of the hallmarks of edge platform.


But What is Edge AI?

It refers to AI algorithms that are processed locally on a hardware device. It is also referred to as on-Device AI. This allows you to process data with the device in less than a few milliseconds, which gives you real-time information. Using Edge AI, one can get personalization features that she wants from the app on the device.

According to IDC, the Edge AI Software market is forecasted to grow from $355 million in 2018 to 1.12 trillion dollars by 2023. Dave McCarthy, research director, IDC, says, “AI is the most common workload in edge computing. As IoT implementations have matured, there has been an increased interest in applying AI at the point of generation for real-time event detection.”


Edge over Cloud

Currently, AI processing is done with deep learning models in a cloud-based data center that require massive computing capacity. And latency is one of the most common issues faced in a cloud environment or IoT devices backed by the cloud. Besides, there is always a risk of data theft or leak during data transfer to the cloud. With edge, data is curated before sending it off to a remote location for further analysis. Further, edge AI shall enable intelligent IoT management.

In edge-based architecture, inference happens locally on a device. This decreases the amount of network traffic flowing back to the cloud with the response time for IoT devices cut to a minimum, thus enabling management decisions to be available on-premise, close to the devices offering numerous advantages.


Drivers Of Edge AI demand:

There are several factors that demand moving AI processing to the edge:

• Real-time customer engagement is irrespective of the user or device location—E.g. Using online payments from device, monitoring exercise activities.
• Ability to run large-scale DNN models on the edge devices. Several frameworks and techniques support model compression, including Google’s TensorFlow Lite, Facebook’s Caffe2Go, Apple’s CoreML, Nervana’s Neural Network Distiller, and SqueezeNet.
• Quick processing and analysis of IoT sensor data.
• Lower bandwidth costs of Edge platforms.


Edge Device Products:

Depending on the AI application and device category, there are several hardware options for performing AI edge processing. The options include central processing units (CPUs), GPUs, application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGA), and system-on-a-chip (SoC) accelerators. The edge, for the most part, refers to the device and does not include network hubs or micro data centers, except in the case of security cameras where network video recorders (NVRs) are included.

The Top 3 edge Products are:

• Nvidia Jetson Nano:
• Intel Neural Compute Stick 2
• Google Edge TPU Dev Board

The most widely used and popular, NVIDIA® Jetson Nano™ Developer Kit delivers computing performance to run modern AI workloads at an unprecedented size, power, and cost. Developers, learners, and makers can now run AI frameworks and models for applications like image classification, object detection, segmentation, and speech processing. It also includes a board support package (BSP), Linux OS, NVIDIA CUDA®, cuDNN, and TensorRT™ software libraries for deep learning, computer vision, GPU computing, multimedia processing, and much more. The software is even available using an easy-to-flash SD card image, making it fast and easy to get started.

Other notable ones include NVIDIA Jetson TX1, TX2, TX2i (which can withstand higher vibration, temperature and humidity ranges, and dust), Sipeed Maixduino Kit for RISC-V AI + IoT, Raspberry Pi 4 Computer Model B, Coral Dev Board, etc.


Real-World Application:

No doubt, edge AI shall be transforming our future. Companies and firms have already started incorporating it to provide an efficient and hassle-free experience to the customers. Some of these instances are:

• Marriott International has partnered with Samsung and Legrand to use IoT and edge AI to create the world’s first IoT-enabled hotel room. The rooms are highly personalized in multiple locations, allowing customers to set up their rooms exactly as desired based on information stored in the app.
• Japanese car manufacturer Toyota is leveraging existing AI edge robotics designed for car manufacturing to assist people with limited mobility.
• Autonomous delivery systems, such as Amazon‘s delivery drones and Domino‘s Robotic Unit, use computer vision to navigate obstacles and optimize routes efficiently. The companies used edge AI to provide data, geo-location, predicted time frame, and personalization for the updates.
• Conversely, an AI-enabled edge computing system in a factory, can contextualize data from multiple machines to detect and ultimately predict problems that cause downtime.
• Expensify‘s virtual assistant, Concierge, assists in the automation of expense reports and travel arrangements for companies. It can inform clients of real-time price changes and can even file receipts on their behalf.


The use and potential of edge AI vary from industry to industry and company to company. Though edge-based inference has proved to be a better alternative to cloud, yet much work needs to be done in this segment. Also, one can try the Intel® Edge AI Fundamentals free course to learn more about this topic.