Yang Chun'an, MIC industry analyst at the Institute for Information Policy, pointed out that generative AI requires a large amount of data and computing power, thus promoting the demand and innovation of software and hardware products such as cloud services, databases, and chips.
(Provided by MIC)
[Reporter Xu Ziling/Taipei Report] The Institute of Industrial Intelligence (MIC) of the Institute of Information Technology (MIC) released the trend of generative AI today (11th) to observe changes in the overall AI investment market. Affected by interest rate hikes, the amount of AI investment in 2022 will decline compared with the previous year However, investment in generative AI has grown against the trend. With the gradual popularization of applications, it is expected to drive investment in the capital market to increase year by year. In 2023, global investment in generative AI will exceed US$10 billion.
Generative AI relies on the underlying infrastructure composed of data and computing power, and superimposed deep learning algorithms on top to construct a generative AI model.
MIC observes technology evolution and points out two major technology trends:
Please read on...
The first is that "special models for vertical fields will flourish." At the current stage, technology development is dominated by general-purpose large-scale models. However, only a few manufacturers are capable of developing the computing power resources and follow-up model inference operations. In the future, vertical fields will need Customizing models to solve specific tasks has seen the emergence of specialized models with reduced parameters across verticals.
The second is that "multimodality will become the basic equipment for general-purpose and special-purpose models." When a single large model can handle multiple tasks such as text, images, videos, and voices, it is called a multi-modal basic model. In the future, whether it is general-purpose ( GPT-4, PaLM-E) or dedicated (Gen 2, DALL·E 2) models, multimodality will be the basic equipment.
Yang Chun'an, MIC industry analyst of the Institute of Information Policy, pointed out that generative AI requires a large amount of data and computing power, so it promotes the demand and innovation of software and hardware products such as cloud services, databases, and chips, and provides emerging applications and business models for various industries. Form a generative AI ecosystem.
Among them, Taiwan's information and communication industry has three major business opportunities: infrastructure and model development companies can use open source resources to develop general models, and provide general large-scale language model API series or consulting services; secondly, use open source resources to develop special models according to user needs Add customized design and provide vertical field-specific models.
The third is the application service provider, who can develop a dedicated model after fine-tuning the model to provide vertical field application services by connecting the APIs of international manufacturers.
Grasp the pulse of the economy with one hand I subscribe to Free Finance Youtube channel
Already added friends, thank you
Welcome to 【Free Finance】
feel good
Already liked it, thank you.
related news