印度憑藉神經形態運算的突破成為全球人工智慧競賽的關鍵參與者

印度快速發展的人工智慧產業也得到了政府的大力支持，政府推出了多項戰略政策

Artificial Intelligence (AI) is rapidly changing the world, becoming an integral part of businesses across various sectors. From healthcare, retail, finance, manufacturing, and supply chain to education, energy, and entertainment industries, there has been a growing demand for AI-powered tools.

人工智慧 (AI) 正迅速改變世界，成為各行業企業不可或缺的一部分。從醫療保健、零售、金融、製造和供應鏈到教育、能源和娛樂產業，對人工智慧工具的需求不斷增長。

According to PWC, AI can transform the productivity and GDP potential of the global economy with the greatest gains to be seen in China and North America, equivalent to a total of $10.7 trillion.

普華永道表示，人工智慧可以改變全球經濟的生產力和 GDP 潛力，其中中國和北美的收益最大，總計相當於 10.7 兆美元。

Driven by this growing usage, the AI market is expected to grow from $50 bln in 2023 to well above $800 bln by the end of this decade, as per Statista.

根據 Statista 的數據，在這種不斷增長的使用量的推動下，人工智慧市場預計將從 2023 年的 500 億美元增長到本十年末的 8,000 億美元以上。

This AI boom has led to competition between large companies to develop the most powerful AI models worldwide, and countries are keen to foster their own competing AI systems.

人工智慧的熱潮引發了大公司之間的競爭，以開發全球最強大的人工智慧模型，各國都熱衷於培育自己的競爭性人工智慧系統。

Amidst this, India has emerged as a key player, with 92% of knowledge workers utilizing generative AI compared to the much lower global average of 75%.

其中，印度已成為關鍵參與者，92% 的知識工作者使用生成式人工智慧，而全球平均則低得多（75%）。

Just late last month, Asia's richest man, Mukesh Ambani, chairman of Reliance Industries, unveiled “JioBrain,” a suite of AI tools and applications to transform businesses in energy, textiles, and more. Reliance's telecommunications business is currently working with the Indian Institute of Technology (IIT) to launch “Bharat GPT” for Indian users.

就在上個月末，亞洲首富、信實工業公司董事長穆克什·安巴尼(Mukesh Ambani) 推出了“JioBrain”，這是一套人工智慧工具和應用程序，旨在改變能源、紡織等領域的業務。 Reliance的電信業務目前正在與印度理工學院（IIT）合作，為印度用戶推出「Bharat GPT」。

“We need to be at the forefront of using data, with AI as an enabler for achieving a quantum jump in productivity and efficiency. “

「我們需要站在數據使用的最前沿，以人工智慧為推動力，實現生產力和效率的飛躍。 “

– Mukesh Ambani

——穆克什‧安巴尼

After building a high-powered IT industry worth $250 bln, India is now setting its eyes on AI services, which, according to a report by Nasscom and BCG, could be worth $17 bln in the next three years.

在建立了價值 2,500 億美元的強大 IT 產業後，印度現在將目光投向人工智慧服務，根據 Nasscom 和 BCG 的報告，未來三年該服務的價值可能達到 170 億美元。

With over 900 million internet users, India has emerged as “the data capital of the world.” The fact that so much data is publicly available is extremely beneficial for companies, as they can write their own AI algorithms.

印度擁有超過 9 億網路用戶，已成為「世界數據之都」。事實上，如此多的數據是公開的，這對公司來說非常有利，因為他們可以編寫自己的人工智慧演算法。

However, computing power and shared resources are needed to accelerate the country's AI industry. For this, the Indian government has procured a thousand GPUs to offer computing capacity to AI makers.

然而，加速國家人工智慧產業發展需要運算能力和共享資源。為此，印度政府購買了一千塊GPU，為人工智慧製造商提供運算能力。

Earlier this year, the first shipment of Nvidia chips arrived in Indian data centers after the CEO of the world's largest chipmaker, Jensen Huang, visited India and had a discussion with Prime Minister Narendra Modi and tech executives.

今年早些時候，全球最大晶片製造商首席執行官黃仁勳訪問印度並與總理納倫德拉·莫迪和科技高管進行討論後，第一批英偉達晶片抵達印度數據中心。

“You have the data, you have the talent. This is going to be one of the largest AI markets in the world.”

「你有數據，你有才華。這將成為世界上最大的人工智慧市場之一。

– Huang told the PM at the time

– 黃當時告訴總理

A Breakthrough: Mimicking the Brain for Smarter Computing

突破：模仿大腦實現更聰明的運算

Amidst all this, scientists at the Centre for Nano Science and Engineering (CeNSe), Indian Institute of Science (IISc), Bangalore, India, made a major breakthrough in neuromorphic computing technology. This technology mimics the human brain's structure and function to create more efficient and intelligent computing systems.

其中，印度班加羅爾印度科學研究所（IISc）奈米科學與工程中心（CeNSe）的科學家在神經形態計算技術方面取得了重大突破。該技術模仿人腦的結構和功能，以創建更有效率、更智慧的計算系統。

This momentous progress can help India become a major player in the global AI race and make AI computing accessible to everyone and integrated into their personal devices.

這項重大進展可以幫助印度成為全球人工智慧競賽的主要參與者，並使人工智慧運算可供每個人使用並整合到他們的個人設備中。

This is certainly a great feat, given that the conventional ‘cloud computing model' requires large data centers that consume a lot of energy. Using resource-intensive data centers limits their use to a small community of developers.

考慮到傳統的「雲端運算模型」需要消耗大量能源的大型資料中心，這無疑是一項偉大的壯舉。使用資源密集型資料中心將其使用限制在小型開發人員社群中。

Neuromorphic hardware promises enhanced energy efficiency and space for AI. However, at present, it can only handle low-accuracy operations. Tasks like NLP, neural network training, and signal processing require substantially more computing resolution and are currently beyond the scope of individual neuromorphic circuit elements.

神經形態硬體有望提高人工智慧的能源效率和空間。但目前只能處理低精度運算。 NLP、神經網路訓練和訊號處理等任務需要更高的計算分辨率，目前超出了單一神經形態電路元件的範圍。

However, the latest advancement by IISc scientists can actually help with it and move the space towards' edge computing,' which moves data processing and storage closer to devices that create and use the data. This reduces latency, improves application performance, and saves network costs.

然而，IISc 科學家的最新進展實際上可以對此有所幫助，並將空間推向“邊緣運算”，從而使資料處理和儲存更接近創建和使用資料的設備。這可以減少延遲、提高應用程式效能並節省網路成本。

Edge computing further enables real-time applications, as well as AI and machine learning applications, to process large volumes of data with greater speed and reliability.

邊緣運算進一步使即時應用程式以及人工智慧和機器學習應用程式能夠以更快的速度和可靠性處理大量數據。

Published in Nature, the latest research by Professor Sreetosh Goswami at the CeNSe, IISC, who led a group of scientists and students, developed a type of semiconductor device called Memristor. Instead of using traditional silicon-based technology, Memristor was created using a metal-organic film.

IISC 領導一群科學家和學生的 CeNSe 教授 Sreetosh Goswami 的最新研究發表在《自然》雜誌上，開發了一種名為憶阻器的半導體裝置。憶阻器不是使用傳統的矽基技術，而是使用金屬有機薄膜製造。

The use of molecular films allowed researchers to track free ionic movements, which widened the memory pathways.

分子薄膜的使用使研究人員能夠追蹤自由離子的運動，從而拓寬了記憶途徑。

To establish kinetic controls over the molecular transition that enabled the neuromorphic traits in a single circuit element, scientists applied voltage pulses and then mapped molecular movements to a distinct electrical signal. This created an extensive ‘molecular diary' of different states.

為了建立對分子轉變的動力學控制，從而在單一電路元件中實現神經形態特徵，科學家施加電壓脈衝，然後將分子運動映射到不同的電訊號。這創造了不同狀態的廣泛“分子日記”。

“Due to this free ionic movement, countless unique memory states and pathways were generated. Such intermediary states had remained inaccessible, so far, as most digital devices are only able to access two either high and low conductance states.”

「由於這種自由離子運動，產生了無數獨特的記憶狀態和路徑。到目前為止，這種中間狀態仍然無法訪問，因為大多數數位設備只能訪問高電導狀態和低電導狀態。

– Professor Sreebrata Goswami, father of Prof Sreetosh

– Sreebrata Goswami 教授，Sreetosh 教授的父親