[자막뉴스] 엔비디아 능가한 '초저전력 AI 반도체'...한국 세계 최초로 개발 / YTN

YTN

6 Mar 202402:32

Summary

TLDRSouth Korean researchers have developed the world's first neuromorphic AI chip that operates at ultra-low power, mimicking the human brain. This breakthrough chip, only 4.5mm in size, can run complex AI models like GPT without external servers, significantly reducing computation time and power consumption. It integrates two AI methods for complementary efficiency, allowing it to perform tasks such as summarizing text, translating languages, and generating responses from learned data. The chip's potential applications are vast, with ongoing research to expand its capabilities beyond language models.

Takeaways

📱 The script discusses the development of next-generation IT devices that can run AI on-device without external servers, emphasizing the challenges due to the high power and size requirements of AI chips.
🧠 South Korean researchers have developed the world's first neuromorphic AI chip that mimics the human brain, capable of running AI with ultra-low power consumption.
🔋 The chip is remarkably small, measuring only 4.5mm in width, and has been successfully tested by running GPT on a computer, demonstrating its ability to perform tasks such as summarizing text and translating languages without external assistance.
⏱️ The chip can process computations that would take a regular computer 45 minutes in just 3 minutes, showcasing its efficiency.
🔍 When compared to Nvidia's GPU A00 graphics card, the new chip consumes only 1/6 of the power and has 1/10 of the area.
📲 The chip has been implemented in a smartphone model called 'Galaxy 24', enabling it to perform language tasks as demonstrated.
🔄 The chip's design incorporates two AI methodologies, deep neural networks for high-accuracy computation with large datasets and spiking neural networks for efficient power usage with smaller datasets, creating a synergistic effect.
💡 The chip operates similarly to the human brain, adjusting its power consumption based on the amount of data it needs to process—using more power for larger tasks and less for smaller ones.
🌐 The achievement of running large language models like GPT on this neuromorphic chip is a world-first, indicating a significant breakthrough in AI and chip technology.
🔬 The research team is currently working on integrating the next version of GPT, GPT 3.0, into the chip and plans to expand their research to various application fields beyond language models.
🎥 The script is from YTN Science, providing a news update on this technological advancement.

Q & A

What is the main challenge in implementing on-device AI for devices like smartphones?
-The main challenge is that AI requires complex computations, which in turn demand a lot of power and can result in larger chip sizes.
What has the Korean research team developed that is a world first?
-The team has developed a neuromimetic AI chip that operates artificial intelligence at ultra-low power consumption, mimicking the human brain.
What is the size of the developed chip?
-The chip is only 4.5mm in width.
How did the team demonstrate the chip's capabilities?
-They mounted the chip on a computer and ran GPT to perform tasks such as summarizing text and translating languages without external server assistance.
What is the performance improvement when the chip is used in comparison to a regular computer?
-The chip can process computations that would take a regular computer 45 minutes in just 3 minutes.
How does the power consumption of the developed chip compare to Nvidia's GPU A00 graphics card?
-The power consumption of the developed chip is only 1/61 of that of the Nvidia GPU A00 graphics card.
What smartphone model was used to implement the chip for language tasks?
-The Samsung Galaxy 24 smartphone model was used to connect the board and implement the language tasks.
What are the two AI methods incorporated in the chip that allow for ultra-low power and small size?
-The chip incorporates deep neural networks for high computation accuracy with large data inputs and spiking neural networks for power efficiency with small data inputs.
How do the two AI methods work together in the chip?
-They work complementarily, with the deep neural network taking over for high-accuracy computations and the spiking neural network handling power-efficient processing for smaller data inputs, creating a synergistic effect.
What is the current focus of the research team in terms of the chip's development?
-The team is currently working on integrating the next version of GPT, GPT 3.0, into the chip and expanding the research scope to various application fields.
What is the significance of this development in the field of AI and semiconductors?
-This development is significant as it marks the world's first successful operation of a large language model like GPT using a neuromimetic, ultra-low power semiconductor, paving the way for more efficient and compact AI solutions.

Outlines

00:00

📱 On-Device AI Chip Development

The script discusses the challenges of implementing on-device AI in next-generation IT devices like smartphones, which require substantial power and size due to complex computation. A Korean research team has developed the world's first neuromorphic AI chip that operates on ultra-low power, mimicking the human brain. This chip, only 4.5mm in size, can perform tasks such as summarizing text, translating languages, and generating responses based on learned data without external server assistance. It achieves what normally takes a computer 45 minutes in just 3 minutes. Compared to Nvidia's GPU A00 graphics card, the chip consumes minimal power and has a compact size. The chip has been implemented in a smartphone model called Galaxy 24, enabling it to handle language tasks efficiently. The achievement is attributed to the integration of two AI methods, a deep neural network for high-accuracy computation with large data inputs and a spiking neural network for power efficiency with smaller data inputs, creating a synergistic effect.

Mindmap

Keywords

💡Smartphone

A smartphone is a handheld device that combines the functionalities of a mobile phone and a computer. In the video, it is mentioned as a device that can potentially run artificial intelligence (AI) without relying on external servers, indicating a shift towards on-device AI capabilities.

💡On-Device AI

On-Device AI refers to the ability of a device, such as a smartphone, to process AI algorithms locally without the need for internet connectivity or external servers. The video discusses the development of AI chips that enable this functionality, emphasizing the importance of reducing dependency on cloud computing for AI tasks.

💡AI Semiconductor

An AI semiconductor, or AI chip, is a type of integrated circuit designed to perform the complex computations required for AI applications efficiently. The script highlights the challenges of developing such chips, which need to handle complex operations while consuming less power and occupying less space.

💡NeuroMorphic AI

NeuroMorphic AI is a concept that involves designing AI systems to mimic the neural structure and functioning of the human brain. The video script introduces a new type of AI semiconductor that operates on this principle, achieving low power consumption and high computational efficiency.

💡Low Power Consumption

Low power consumption is a critical feature for AI chips, especially in portable devices, as it extends battery life and reduces heat generation. The script mentions that the newly developed AI chip consumes significantly less power than traditional GPUs, which is a major advancement in AI technology.

💡Deep Neural Networks (DNN)

Deep Neural Networks are a subset of machine learning algorithms that are composed of multiple layers of artificial neural networks. They are known for their ability to process large amounts of data and perform complex tasks. The script refers to DNNs as the traditional AI method that provides high computational accuracy when dealing with large datasets.

💡Spiking Neural Networks (SNN)

Spiking Neural Networks are a type of neural network that more closely resembles the biological neural networks of the human brain. They are known for their power efficiency and are highlighted in the script as the technology that enables the new AI chip to operate with low power consumption when processing smaller amounts of data.

💡Synergistic Effect

The synergistic effect refers to the phenomenon where the combined effect of two or more elements is greater than the sum of their individual effects. In the context of the video, the combination of DNNs and SNNs in the AI chip creates a synergistic effect, enhancing both computational accuracy and power efficiency.

💡GPT (Generative Pre-trained Transformer)

GPT is a type of AI language model that is trained on large datasets to generate human-like text. The script mentions that the new AI chip was able to run GPT, demonstrating its capability to handle complex language processing tasks without external server support.

💡GPT 3.0

GPT 3.0 is the next version of the GPT model, which is expected to have improved capabilities over its predecessors. The script indicates that the research team is working on integrating GPT 3.0 into the AI chip, suggesting a continuous advancement in the field of AI language models.

💡Galaxy 24

Galaxy 24 is mentioned in the script as a smartphone model where the new AI chip has been implemented. This showcases the practical application of the chip in a consumer product, enabling the phone to perform advanced AI tasks such as language translation and data summarization.

Highlights

Smartphones and other devices can now run AI without external servers, thanks to on-device AI security.

Implementing this technology is not easy due to the complex computations and power requirements of AI chips.

A Korean research team has developed the world's first neuromorphic AI chip that operates on ultra-low power, mimicking the human brain.

The chip, measuring only 4.5mm in size, has been successfully tested by running GPT on a computer.

It can perform tasks such as summarizing text and translating languages without external server assistance.

The chip can process computations that normally take 45 minutes on a regular computer in just 3 minutes.

Compared to Nvidia's GPU A00 graphics card, the new chip consumes only 1/6 of the power and has a smaller footprint.

The chip has been implemented in a smartphone model called Galaxy 24, enabling it to perform language tasks.

The chip incorporates two AI methods, leveraging their respective strengths for optimal performance.

Deep neural networks are used for high-accuracy computations with large datasets.

Spiking neural networks are employed for their power efficiency with smaller datasets.

The synergistic effect of these two technologies allows for ultra-low power consumption and high efficiency.

The chip mimics human brain activity, consuming more energy with complex thoughts and less with simpler ones.

This is the first time in the world that a neuromorphic chip has successfully run a large language model like GPT.

The research team is currently working on integrating GPT 3.0, the next version of the language model, into the chip.

They plan to expand the research scope to various application fields beyond language models.