Track Introduction
Teams participating in this track are tasked with training their own models to tackle challenges within the Honor of Kings scenario. The emphasis is on developing solutions for both single-agent and multi-agent systems, as well as investigating the generalizability and adaptability of models.
Target Participants
This track is open to full-time students from global higher education institutions who are currently enrolled in associate degree, undergraduate, master's, or doctoral programs during the registration period.
Participants must compete as part of a team. Each team member must be from the same university, with a maximum of 5 members per team, including at least one faculty advisor. Team members can be from any field of study, and each individual is allowed to join only one team.
This track includes both a "Domestic Region" and an "International Region"（including Hong Kong, Macao, and Taiwan area"）,with competing teams required to choose their region according to the geographical location of their university. Teams from universities in the Sichuan-Chongqing area of China will be ranked separately during a specific segment of the competition. For detailed information on matchmaking and the rules for progressing through the competition, please see the section below on this page.
Format Intro
Rewards
The prize money for this track is set as above, with amounts in Chinese Yuan (CNY) and pre-tax.
In addition to the prize money, participating teams will also receive the following rewards:
1.All teams completing the competition (by successfully submitting a model and passing verification) will be awarded an official certificate of participation.
2.Exceptional participants may have the opportunity to secure internships related to Tencent's AI Arena Program and gain priority access to Tencent Group's campus recruitment and internship programs. Detailed information will be provided later.
Schedule
About the Challenge：In this challenge, participating teams need to use algorithms to train models that drive agents to learn movement strategies through continuous exploration of the map. They must wisely use summoner skills and acceleration boosts to reach the endpoint within a set time frame while collecting as many treasure chests as possible.
The map features starting and ending points, roads, obstacles, acceleration boosts, and treasure chests. Agents have a local field of view and can move around the map, deploy summoner skills, and acquire rewards contained within the treasure chests. (A comprehensive development guide with more details will be provided on the platform following successful registration.)
Objective：Participating teams must locally train and submit a model within a given timeframe. Their goal is to control agents on the assessment map to gather as many scores as possible in the least amount of time, fulfilling the mission of embarking on a return adventure to the mystic realm.
Ranking Rules： After the submission phase ends, the system will automatically execute the challenge with the latest model submitted by each participating team. Teams will be ranked according to their scores, and these rankings will constitute the final results for this phase of the competition.
Advancement Rules： The top 130 teams on the scoreboard of "Domestic Region" (including the top 30 teams from the Sichuan-Chongqing scoreboard) will advance to the next stage.
The top 30 teams on the scoreboard of "International Region" will advance to the next stage.
About the Challenge：In this challenge, participating teams are required to train model-driven agents using algorithms. These agents will continuously explore the 1v1 map of the game "Honor of Kings" to learn the optimal strategy, aiming to be the first to destroy the defensive tower in front of the opponent's camp crystal to achieve victory.
The map used in this challenge is elongated, with resurrection points for both teams' agents at each end, and a crystal belonging to each camp positioned in front of the resurrection points. These crystals continuously produce minions for their respective camps, which automatically advance towards the opponent's camp, attacking defense towers, crystals, and heroes along the way. Ahead of the crystals are the camp's defense towers, capable of attacking enemy heroes and minions within range. Agents have the freedom to move around the map and unleash skills at will. (Detailed instructions will be provided in the development guide available on the platform after successful registration)
Objective：Teams are required to utilize the allocated computing resources within a specified timeframe to train their models. The objective is for these models to learn the optimal winning strategy through continuous exploration of the 1v1 maps, enabling them to secure as many victories as possible in matches against other teams.
Ranking Rules：At the conclusion of the tournament submission phase, the system will automatically conduct multiple rounds of matchmaking using the latest model submitted by participating teams. Each team will engage in an equal number of matches against all opponents in their respective division and will be ranked based on their accumulated points. During each round of matchmaking between two teams, all heroes from both sides must participate, facing off against each of the opposing team's heroes in two consecutive rounds. Winning a round earns 1 point, while losses yield no points.
Advancement Rules： The top 8 teams on the scoreboard of "Domestic Region" will progress to the next stage. The top 8 teams on the scoreboard of "International Region" will advance to the next stage.
About the Challenge：In this challenge, teams train models to control 2 agents working alongside 1 platform-bot teammate. Their goal is to master the best strategy on the Honor of Kings 3v3 map, aiming to destroy the opponent's base crystal first to win.
The map features a main lane flanked by upper and lower jungle areas. Each team's spawn points are at the map's ends, with their base crystals spawning minions that advance towards the enemy, attacking towers, crystals, and heroes en route. Defense towers protect the crystal, targeting enemy heroes and minions within their range. The jungle areas are key for gaining gold and experience, shared between both teams but with camp-specific entry restrictions. (Hint: Use these to escape enemies.)
Agents can freely move and use skills on the map. Over time, they earn gold and experience, with significant rewards for killing jungle monsters, enemy minions, heroes, or towers. The team that destroys the enemy crystal claims victory.
Objective：Participating teams must train their models within a set timeframe using allocated computational resources. They are to work with platform-bot teammates to constantly explore the 3v3 maps, aiming to discover the most effective strategies for collaborative victory. The objective is to secure the highest number of wins in matches against other teams.
Ranking Rules：After the submission phase ends, the system will automatically run multiple rounds of battles using the most recently submitted models from the participating teams. In this phase, teams will face off in an equal number of rounds against all other teams, with rankings based on the points each team accumulates.
In each match-up between two teams, all heroes must be deployed and matched against each opposing team's heroes in turn, leading to two matches with teams swapping sides. A win awards 1 point, while a loss awards no points.
Recommendations
It is advised that each team possess at least one computer adhering to the recommended specifications below to establish the local environment for development and training.