The AI ​​Generation Race

SGGP 14/09/2023 14:09

Generative artificial intelligence (AI) – a type of AI capable of generating new content such as text, images, videos , and audio – is being widely and extensively used to create new products and services such as chatbots, virtual assistants, and creative tools, with applications in many areas of life.

Professor Vu Ha Van, Director of Science at VinBigdata, introduces AI generation.

Many applications

Leading global technology companies are developing generative AI at a fairly rapid pace. Google, in particular, has pioneered the development of several generative AI systems, such as Deep Dream (which can create abstract images from real-world photos), Dal-e (which can create images from text descriptions), and GPT-3 (which can generate text, translate languages, and write various types of creative content).

Recently, Google announced new product enhancements to encourage businesses and public sector organizations in Southeast Asia to experiment with and build applications using large-scale language models and generative AI. Accordingly, Google's Vertex AI, as a comprehensive AI platform, allows customers to access, customize, and deploy models, as well as build and scale generative AI applications.

Amazon has developed several generative AI systems, such as Amazon Rekognition (which analyzes images and videos to detect and recognize objects) and Amazon Lex (a chatbot service that can communicate and interact with humans like a human). Meanwhile, Facebook has been involved in generative AI since early on, developing several generative AI systems, notably DeepText (a generative AI system that can translate languages, write various types of creative content, and answer user questions comprehensively).

AI generation in a "packaged" form for sale has also emerged, such as VMware Technology Group's recent launch of the Private AI service package to encourage businesses to apply AI generation. “VMware Private AI realizes a future vision where AI generation serves every member of the enterprise by bringing the choice of computing models and AI technologies closer to the enterprise's data. Private AI offers many use cases, from software development and marketing content creation to customer service tasks and data analytics,” said Chris Wolf, Vice President of VMware AI Labs.

Vietnam joins the fray.

In Vietnam, leading technology companies have developed generative AI for various applications. Specifically, FPT Software has developed several generative AI systems such as: FPT AI Writer (a generative AI system capable of writing text, translating languages, and answering user questions comprehensively); and FPT AI Chatbot (capable of communicating and interacting with humans like a human, currently used in applications such as customer service and sales). The AI ​​startup Lumitics is developing a generative AI system that can create realistic images, videos, and audio. This system is being used in applications such as entertainment, education, and marketing.

Professor Thoai Nam from the University of Technology (Vietnam National University Ho Chi Minh City) stated that generative AI creates many valuable applications, realizing AI applications in daily life. However, generative AI can also bring some risks, such as bias due to its training data, and the potential for using it to create fake creative content. This could negatively impact trust in the application of the technology, so we need thorough preparation in terms of both technological foundations and legal policies to mitigate negative impacts and promote the positive exploitation and application of generative AI.

According to an announcement by VinBigdata Joint Stock Company, the company has successfully built a large-scale Vietnamese language model, laying the foundation for developing AI-generated solutions. The complete mastery of the technology, self-development from the initial stages, and successful construction of the large-scale Vietnamese language model is considered a significant step forward, enabling VinBigdata to integrate AI-generated technology into its ecosystem of products and services already offered to the market.

VinBigdata will integrate technology to make VinBase (a comprehensive multi-cognitive AI platform) the first generative AI platform in Vietnam; and simultaneously provide development solutions based on generative AI technology such as Generative AI chatbots, callbots, or the next-generation ViVi virtual assistant… helping to increase the naturalness of machine communication, supporting users to search and synthesize information much faster and simpler than before.

Professor Vu Ha Van, Director of Science at VinBigdata, shared: “In Vietnam, VinBigdata has successfully built a Vietnamese big language model, focusing on solving three core problems: improving accuracy, minimizing computing infrastructure costs, and ensuring security. Instead of needing approximately 175 billion parameters like ChatGPT, VinBigdata can create a big language model with a few billion parameters while still being able to generate highly authentic texts, focusing on Vietnamese data and knowledge.”

Source: https://www.sggp.org.vn/cuoc-dua-ai-tao-sinh-post705529.html