Sunday, September 25

What is edge AI and how does it work?

As part of the preparations for GTC 2022, NVIDIA has published an article that addresses the new opportunities that have arisen for edge AI, which were previously unimaginable: “from helping radiologists identify pathologies in the hospital, to driving cars down the highway, through helping us pollinate plants.”

Countless analysts and companies are talking about and implementing edge computing, whose origins date back to the 1990s, when content delivery networks were created to serve web and video content from edge servers deployed near users.

Today, almost every company has job functions that can benefit from adopting edge AI. In fact, edge applications are driving the next wave of AI to improve people’s daily lives at home, at work, at school, and on the go.

What is edge AI?

AI at the edge is the deployment of AI applications to devices throughout the physical world. It’s called “edge AI” because the AI ​​computation is done close to the user at the edge of the network, close to where the data is, rather than centrally in a cloud computing facility or on a remote server. private data center.

Since the Internet has a global reach, the edge of the network can connote anywhere. It can be a store, a factory, a hospital, or the devices that surround us, such as traffic lights, autonomous machines, and telephones.

AI at the edge: why now?

Organizations across all industries are looking to increase automation to improve processes, efficiency and security.

To help them, computer programs need to recognize patterns and execute tasks repeatedly and safely. But the world is unstructured, and the range of tasks humans perform encompasses infinite circumstances that are impossible to fully describe in programs and rules.

Advances in cutting-edge AI have opened up opportunities for machines and devices, wherever they are, to work with the “intelligence” of human cognition. AI-enabled smart apps seek to perform similar tasks under different circumstances, much like real life.

The effectiveness of deploying AI models at the edge stems from three recent innovations.

1. The maturation of neural networks: Neural networks and related AI infrastructure have finally developed to the point of enabling pervasive machine learning. Organizations are learning how to successfully train AI models and deploy them to production at the edge.

2. Advances in computing infrastructure: Powerful distributed computing power is needed to run AI at the edge. Recent advances in highly parallel GPUs have been adapted to run neural networks.

Also Read  Calviño se queda sentada y no se suma al boicot de EE.UU. a Rusia en el G-20

3. Adoption of IoT devices: The widespread adoption of the Internet of Things has fueled the explosion of big data. With the sudden ability to collect data on all aspects of an enterprise—from industrial sensors, smart cameras, and robots to name a few—we now have the data and devices to deploy AI models at the edge. Additionally, 5G is giving the IoT a boost with faster, more stable, and more secure connectivity.

Why deploy AI at the edge? What are the benefits of AI at the edge?

Since AI algorithms are capable of understanding language, images, sounds, smells, temperature, faces, and other analog forms of unstructured information, they are especially useful in places occupied by end users with real-world problems. . These AI applications would be impractical or even impossible to deploy in a centralized cloud or enterprise data center due to latency, bandwidth, and privacy issues.

Edge AI benefits include

– Intelligence: AI applications are more powerful and flexible than conventional applications that can only respond to inputs that the programmer intended. In contrast, an AI neural network is not trained to answer a specific question, but to answer a particular type of question, even if the question itself is new. Without AI, applications would not be able to process infinitely diverse input such as text, spoken words, or video.

– Real-time information: Because edge technology analyzes data locally rather than in a distant cloud delayed by long-distance communications, it responds to user needs in real time.

– Reduced cost: By bringing processing power closer to the edge, applications need less internet bandwidth, greatly reducing network costs.

– Greater privacy: AI can analyze real-world information without ever exposing it to a human, greatly increasing the privacy of anyone whose appearance, voice, medical image, or other personal information needs to be analyzed. Edge AI further enhances privacy by holding that data locally, uploading only the analytics and insights to the cloud. Even if some of the data is uploaded for training purposes, it may be anonymized to protect the identity of users. By preserving privacy, edge AI simplifies the challenges associated with data compliance.

– High availability: Decentralization and offline capabilities make edge AI more robust as internet access is not required to process data. This translates into higher availability and reliability for production-grade and mission-critical AI applications.

Also Read  El temor a un repunte de casos de Covid lleva a las comunidades a recomendar la mascarilla en ciertos espacios

– Steady improvement: AI models are getting more accurate as they are trained on more data. When an edge AI application is faced with data it can’t process accurately or reliably, it often uploads it so the AI ​​can retrain and learn from it. Therefore, the longer a model is in production at the edge, the more accurate the model will be.

How does AI technology work at the edge?

For machines to see, detect objects, drive cars, understand speech, talk, walk, or in any way emulate human abilities, they need to functionally replicate human intelligence.

AI uses a data structure called a deep neural network to replicate human cognition. These deep neural networks are trained to answer certain types of questions by showing them many examples of those types of questions along with the correct answers.

Known as “deep learning,” this training process is typically run in a data center or cloud due to the large amount of data required to train an accurate model, and the need for data scientists to collaborate on setup. of the model. After training, the model graduates to become an “inference engine” that can answer real-world questions.

In edge AI deployments, the inference engine runs on some kind of computer or device in faraway places like factories, hospitals, cars, satellites, and homes. When AI encounters a problem, the problematic data is often uploaded to the cloud to further train the original AI model, which eventually replaces the inference engine at the edge. This feedback loop plays an important role in increasing the performance of the model; once edge AI models are deployed, they only get smarter and smarter.

What are examples of use cases for AI at the edge?

AI is the most powerful technological force of our time. We are at a time when AI is revolutionizing the world’s largest industries.

In manufacturing, healthcare, financial services, transportation, energy, and more, cutting-edge AI is driving new business outcomes across industries, including

– Intelligent forecast in energy: For critical industries like energy, where discontinuous supply can threaten the health and well-being of the general population, smart forecasting is key. Edge AI models help combine historical data, weather patterns, grid status, and other information to create complex simulations that inform more efficient generation, distribution, and management of energy resources for customers.

Also Read  Díaz buscará independientes que desplacen a Podemos de los primeros puestos en las listas

– Predictive maintenance in manufacturing: Sensor data can be used to detect anomalies early and predict when a machine will fail. Equipment sensors scan for faults and alert management if a machine needs repair so the problem can be addressed early, avoiding costly downtime.

– Instruments with AI in healthcare: Modern medical instruments at the edge are becoming AI-enabled devices that use ultra-low latency surgical video streaming to enable minimally invasive surgeries and on-demand insights.

– Intelligent virtual assistants in retail: Retailers are looking to improve the digital customer experience by introducing voice ordering to replace text-based searches with voice commands. With voice ordering, shoppers can easily search for items, request product information, and place orders online using smart speakers or other smart mobile devices.

What role does cloud computing play in edge computing?

AI applications can run in a data center like those in public clouds, or at the edge of the network, close to the user. Both cloud computing and edge computing offer advantages that can be combined when deploying edge AI.

The cloud offers benefits related to infrastructure cost, scalability, high utilization, resilience to server failure, and collaboration. Edge computing offers faster response times, lower bandwidth costs, and resiliency to network failures.

Cloud computing can support the deployment of AI at the edge of the network in several ways:

  • The cloud can run the model during its training period.
  • The cloud continues to run the model as it is retrained on data coming from the edge.
  • The cloud can run AI inference engines that complement models in the field when high computing power is more important than response time. For example, a voice assistant can respond on your behalf, but send complex requests to the cloud for analysis.
  • The cloud serves the latest versions of the AI ​​model and application.
  • The same edge AI often runs across a fleet of devices in the field with cloud software

NVIDIA announces multiple news, and more than 900 online sessions with industry experts for GTC 2022. Those interested in attending can register for free at this link.

Leave a Reply

Your email address will not be published.