Although several artificial intelligence (AI) companies in the early days put forward AI platform businesses, it has been found that vertical AI companies that reduce high barriers in the platform business and solve problems in various industries are emerging one after another.
Yoon Jong-young, head of AI Yangjae Hub Center, delivered a presentation on the theme of "Domestic AI Trends and Directions" at the "August Startup Oasis" webinar co-hosted by the Seoul Creative Economy Innovation Center and AI Yangjae Hub on the 26th. Yoon said, "In the early days of the AI market, chatbot companies based on natural language processing were at the center and more companies were trying to solve problems based on computer vision in a specific area. Since then, more AI technology companies are specializing in industries."
Yoon said, "Recently, the Generation Pre-traination Transformer (GPT-3) of OpenAI, a non-profit organization in the United States, is drawing keen attention as an issue beyond AI. Competition is taking place to see who performs better between GPT-3 and humans."
GPT-3 is an AI engine that allows you to talk to people using sentences from movies and even code, and VWO, a startup, used it to make better headlines, lists, and click buttons on its website and compete with AI to bring in more traffic. It became a topic of discussion by running a competition between AI and humans to see who could make it better and get more traffic.
He also said, "With the combination of computer power to cognitive, creative, and learning skills, new AI services are emerging, and not only GAN (generative hostile neural networks) but also Auto ML (automated machine learning) that can replace developers in the future." "AI companies are transforming their businesses to provide better value with higher accuracy, lower price, and higher quality," he said.
Citing a paper, he said that the AI market is developing in fields such as, conversational AI, computer vision, data annotation, robotics, and autonomous. Conversational AI enables services such as remote medical treatment, robotics that evolve from social robots to collaborative robots that care for people, and autonomous technologies such as automobiles capable of autonomous driving are emerging. Conversational AI enables services such as remote medical treatment, robotics that evolve from social robots to collaborative robots that care for people, and autonomous technologies such as automobiles capable of autonomous driving.
In a presentation on the topic 'The Current State of Natural Language Processing Technology', Sun-ho Baek, CEO of AI natural language processing technology company, said, "Many people expected that machine translation would replace humans with the advent of Google's neural network-based translator, but the neural network-based translator training data has limitations as it is general-purpose data, so machine translation has yet to keep up with expert translation.”
He also explained that natural language processing technology is developing centered on user data, and AI is changing from data centric to customer centric. CEO Baek said, “In the past, there was a perception that companies with high-quality data made good translators for the language, but now, only high-quality data is insufficient, and only companies that provide services that meet user needs can survive.”
[Reporter Yoon-Jeong Chae]