The final round, titled "Mine Golds, Unfold Knowledge," will see the top eight teams go head to head across 15 rounds, using artificial intelligence (AI) algorithms and simulating an esport tournament format.

Each team will create an agent, an autonomous entity that can learn and make decisions strategically to defeat its opponents based on the Reinforcement Learning (RL) algorithm.

One representative from DeepShuttling, a team of Vietnamese engineers from Singapore, said: "Reinforcement learning is a new and exciting field. It is gaining popularity not only in Vietnam but also around the world. Despite limited knowledge, over the past two months, the group has gradually trained its 'gold miners' to become smarter thanks to previous match experience and expert suggestions."

After nearly two months, it has become an intellectual event between humans and AI where machine learning techniques allow agents to make smarter decisions via self-study and experience accumulation. The effectiveness of the RL algorithm has been proven by the fact AI can beat real people in the most difficult games today, from chess to esports like Dota2 and Starcraft II.

After the group stage, participants must continue to update their programming code with consultation from AI experts, including machine learning engineers Nguyen Xuan Phong from Canada's Mila (Montreal Institute for Learning Algorithms) and Vu Huu Tiep from Google.

"Myself and the organizers are quite surprised by the quality of contestants in this competition. The algorithm is difficult to grasp, while the problem posed by the organizing committee is not easy. Teams competed fiercely to prove their ability, creating an attractive group stage," Phong said.

Around 1,000 Vietnamese engineers around the world registered for the tournament. The first place winner will receive a prize of VND100 million ($4,300) and a workshop course worth VND20 million from global IT services and outsourcing company FPT Software.