SKT said on the 25th that its research paper on an AI recommendation model was invited to the global AI conference AAAI 2026, being held at Singapore Expo, and was presented on stage on the 24th.
AAAI is an AI conference regarded as among the world's top, along with the Neural Information Processing Systems conference (NeurIPS). This year, about 24,000 papers were submitted, up from about 13,000 the previous year, while the acceptance rate fell to 18% from 23%, intensifying competition.
SKT's paper was selected for an on-site presentation, an opportunity given to only about the top 4% of all submissions, and also conducted a poster presentation as part of the conference program.
The paper presented this time covers research on version 4.0 of One Model, an LLM-based AI recommendation model independently developed by SKT. It also introduced a method for applying "reasoning Reinforcement Learning," which has recently drawn attention among AI researchers worldwide, to real service environments.
One Model version 4.0 is an AI recommendation model that comprehensively understands various behavioral data such as customers' clicks, usage histories, and interests, and then generates in natural language: ▲ which products and services to recommend ▲ why the recommendation was made ▲ how to compose the marketing message to deliver to the customer.