Kibo, New Technology Evaluation System 'Airate' Fully Applied to Evaluation Practice

Improvement of Over 9% in Predictive Performance of Evaluation Models Through Introduction of Artificial Intelligence Algorithms

[Asia Economy Reporter Kim Cheol-hyun] The Korea Technology Finance Corporation (Chairman Jung Yoon-mo, hereinafter Kibo) announced on the 20th that it has completed the development of 'AIRATE,' an AI-based new technology evaluation system, based on a comprehensive diagnosis of the previously operated technology evaluation model (KTRS), and has fully applied it to technology evaluation tasks since January this year.

Kibo developed a system that evaluates the value of patents by training AI with the evaluation patterns of expert evaluators, based on proactive research for the financial application of AI technology, which has been actively discussed recently. Based on this, AI algorithms were applied to the technology evaluation rating model that grades the value of technology.

Kibo reorganized the technology evaluation model, which had been operated based on statistical models, into an AI-based standard model system and launched the brand name 'AIRATE' to enhance recognition of the distinctiveness and excellence of the new technology evaluation system. It means a collaboration system where experts and AI create mutual synergy. AIRATE dramatically improved the predictive performance of the model. By introducing AI algorithms, the prediction accuracy improved by more than 6% for technology business growth potential and more than 12% for technology business risk potential compared to the existing model, resulting in an overall increase of more than 9%.

Until now, early startups and venture companies with high growth potential but weak finances found it difficult to receive additional funding. However, through AIRATE, it is now possible to provide broader support to technology SMEs with sufficient growth potential even if their finances are weak. Not only the technology business evaluation rating but also the technology business growth rating and technology business risk rating can be used separately, allowing these ratings to be independently calculated and combined in various ways to design products tailored to technology SMEs with expected growth.

AIRATE resolved the model bias that can occur when only accuracy is pursued by implementing a competitive and cooperative evaluation system between experts and AI. Additionally, by securing high-quality data that AI can continuously learn and evolve from, and introducing interpretable and explainable AI into the model, it is evaluated as an innovative case that overcomes the limitations pointed out as drawbacks of AI in the financial sector.

Chairman Jung Yoon-mo of Kibo said, "Kibo's new technology evaluation system is a case that officially introduced AI techniques into technology evaluation, and it can be called a collaboration model where expert insight and stability and AI accuracy cooperate with each other," adding, "We will strengthen the competitive evaluation system and proactively respond to the fast-follower strategies of competing institutions."

This content was produced with the assistance of AI translation services.