AI everywhere: across the edge and sustainably

The integration of artificial intelligence (AI) is everywhere, providing transformative opportunities for various industries.

One such paradigm shift is the convergence of artificial intelligence and edge computing, promoting sustainable solutions and innovative applications.

Enterprises can leverage the rapid development of artificial intelligence to enable hyper-personalization at scale in customer experience (CX) and predictive analytics to transform their services and operations to manage business operations.

Integration of Artificial Intelligence, 5G and IoT

The benefits provided by 5G networks are:

• Low latency
• Significant increase in device connections This in turn allows for the expansion of machines to enable machine communication
• The emergence of network-connected devices and sensors has led to massive hyper-personalization

Environment and Economy

Popular Waterhouse Coopers has released a report setting out the potential of artificial intelligence to help reduce carbon emissions. According to its analysis, by 2030, the artificial intelligence economy applied in the four major fields of agriculture, energy, transportation and water may bring up to:

• contribute to global economic growth of US$5.2 trillion or GDP 4.4%.
• Greenhouse gas (ghg) emissions were reduced by 240 million tons, or 4%.
• Global net employment increased by 38.2 million, or 1%.

In this way, environmental goals and economic goals can be coordinated with each other, especially through technological progress. As businesses and the overall economy grow, more efficient AI is more effective at a macroeconomic and social level, able to scale and create economic and job growth. At a microeconomic level, by lowering the cost of deploying and scaling AI, businesses may expand into new services, products, and business models, and enable startups to thrive and scale. At the same time, achieving this with lower energy consumption reduces the carbon footprint.

In addition, a group of leading scientists in the field of artificial intelligence mentioned that machine learning can be used to assist in combating climate change, across electrical systems, industry, transportation, construction, smart grids, disaster management and other industries. These challenges ensure the importance of scaling up AI on an efficient basis that combines cost and environmental benefits. Energy efficiency is key in both areas.

The emergence of production artificial intelligence has caused a craze, which is often provided by large language models (LLM). These models employ transformers and self-attention mechanisms, often combined with deep reinforcement learning, to optimize their responses. While these models are computationally expensive, including hardware requirements, energy costs, and carbon footprints, their inclusive service requirements, energy costs, and carbon footprints are also reduced.

AI is everywhere: from intelligence to the “smart” edge

Smart refers to devices that are connected to the Internet. However, connected devices are becoming increasingly “smarter” as AI is embedded locally on devices, such as PCs with AI. In this case, intelligence refers to the ability to respond meaningfully to the user and personalize the experience, rather than human-level intelligence.

As the Internet of Things scales, the growth of edge computing will require ultra-low latency, which in turn allows for real-time responses.

As mentioned above, artificial intelligence will increasingly be at the edge of the network - called edge computing or simply edge, where the processing of data is closer to where it is generated and may actually be located on the device itself. This keeps latency very low, resulting in real-time responses to users.

Cloud/edge hybrid with security and reliability as key factors

The cloud model will continue to be applied to data centers, providing important resources and capabilities for storing historical data for analysis. This will also allow for ongoing algorithm development using hybrid models, supporting training of AI models on cloud servers and inference of AI at the edge, providing further potential for personalization at scale.

Examples of Edge Artificial Intelligence

• Smart grids enable real-time two-way information flow and combine this with AI models such as Google DeepMind’s NowCast and GraphCast to predict weather and optimize renewable energy Supply and demand management.
• Microgrids powered by IoT can operate in grid-tied or standalone island settings and enable locally produced energy, manage outages and increase efficiency.
• Smart meters with built-in sensors can transmit real-time information, detect power outages and monitor power supply quality.
• Battery optimization for renewable energy storage.
• Drones with computer vision can inspect solar panels and wind turbines and detect damage, thereby reducing power generation.
• Unexpected outage prediction and automatic intervention.
• The development of green hydrogen and fuel cells.
• Green artificial intelligence for automated machine learning.
• Urban traffic management planning, predicting traffic congestion and rerouting traffic.
• Design algorithms for electric vehicle operation to optimize the relationship between battery charging, distance and available charging points.
• Artificial intelligence has been deployed in the construction of smart buildings, and IoT sensors can detect whether there are people in a room and adjust heating/air conditioning or lighting accordingly to optimize energy consumption.
• Apply generative artificial intelligence to the construction and planning phases of a building to predict potential issues with digital twins and optimize sustainable designs.
• In the manufacturing sector, predictive analytics is applied to unplanned downtime and automation to reduce their occurrence, thereby optimizing production operations and reducing the waste that such downtime can cause.
• Optimize energy consumption and carbon footprint of manufacturing processes and supply chains.
• Recommendations for the retail industry are applied alongside predictive analytics, allowing brands to enhance demand forecasts and optimize their supply inventory and production.

The above is the detailed content of AI everywhere: across the edge and sustainably. For more information, please follow other related articles on the PHP Chinese website!