Basic Information
One sentence introduction: AI Arena is a PVP fighting game where players can learn through AI and continuously evolve their game characters. The battle mode is similar to Nintendo’s Super Smash Bros.
Features: The worlds first fast-fighting fighting game that combines AI and NFT. Players train AI to improve character abilities, adjust the characters numerical distribution through the observation room, and then put them into battle in the arena to gain a higher ranking.
Vision: Combining games, NFT (IP) and AI algorithms to bring players a new gaming experience while protecting the intellectual property rights of AI practitioners and helping them monetize. The long-term vision is to create an AI-powered two-sided market that matches the needs of buyers and sellers.
NFT asset performance: AI Arena Mint Pass - 4.11 ETH, an increase of 215% in the past 30 days (deadline: 2024.01.09 data source https://opensea.io/collection/ai-arena-mint-pass)
Game DEMO: https://www.youtube.com/watch?v=xcXjY55xUVYt=2s
Investors: Paradigm; Framework Ventures; SevenX Ventures, etc.
Interpretation Project
Cross-narrative of AI+Web3
In a previously published research article on ZKML, SevenX mentioned that decentralized blockchain is an important force in balancing AI (and machine learning). First, it can be combined with other technologies, such as ZK, to optimize trust in machine learning. The second is that it can effectively utilize long-tail resources and reduce the cost and threshold of using AI. On the other hand, because many Web3 applications sacrifice user experience for the sake of security and decentralization, AI can help optimize and improve User experience, this is the part where AI can empower Web3
Specific to the application scenarios, although AI+DeFi and AI+DID/social all have use cases, the combination of AI and Web3 games may be the first area to expand the user base.
Original gameplay innovates user experience
AI Arena is not just a Web3.0 game that deeply applies AI. Instead, we want to allow more users to contact, understand and experience AI through Web3.0 games, and allow artificial intelligence engineers to provide a variety of AI based on AI Arena. AI algorithms, thereby helping them increase their revenue.
Players first create their own AI character (the character does not need to be NFT, NFT characters can receive rewards. In addition to purchasing NFT on the Marketplace, players can also win NFT by training the AI to participate in high-quality competitions), and then pass the training mode Start training for your role. Players can choose a training venue of their choice to fight (train) with robots, and teach their AI characters how to fight through combat.
(When data reception is turned off)
(After data reception is turned on)
During combat, the AI characters appearance will be covered with red lines, indicating that the players AI character has not started receiving data (the character has not been trained). Therefore, players should turn on data reception before officially training the AI to ensure that the AI can receive data. At this time, the appearance of the AI will be covered with blue lines (the reason for this design is to prevent the AI from receiving unnecessary combat data. As a result, the trained AI cannot achieve the ideal combat effect)
After completing data collection, players need to set the training parameters of the AI, including: training intensity (number of single training iterations, block size, learning rate, direction function, action function), focus areas and data that needs to be cleared, and then start training the AI Role
AI Observation Room: After training is completed, players can check the training effect through the AI Observation Room, and can choose to save the model or abandon the model. If the player chooses to save the training model, a metamask pop-up window will prompt. After confirmation, the model will be saved on IPFS and the hash will be stored on the blockchain.
After the training is completed, players can play against the system robot in simulation mode to test their training results.
“Fusion Pool” – AI Arena’s Economic Flywheel
AI Arena’s economic growth flywheel is driven by Tokne and NFT together. Mainly involves two types of users: players and researchers
player
Stake $NRN to participate in research competitions (AI algorithm models) and have the opportunity to receive more $NRN rewards
Players train AI characters to participate in game rankings. To participate, you need to stake $NRN (equivalent to entry fee) and have the opportunity to receive more $NRN rewards (AI Arena will distribute rewards regularly). Players who fail to win will lose part of their bet
Before entering a ranked round, players can choose to invest a percentage of their $NRN winnings from that round into the Fusion Pool. The more bonuses transferred to the Fusion Pool, the higher the probability of winning NFTs.
researcher
Players can use $NRN to buy batteries, charge NFTs and then battle in ranked matches
Merging Pool
The main function of the fusion pool: giving players a chance to obtain more powerful NFT characters
The system will select all NFTs that were staked in a specific round of ranked play and let them compete against each other. Through competition, the system can select which NFTs to fuse to create more powerful NFTs. The system uses a fitness function to score the strength of the NFT. After obtaining the fitness score of each NFT, the system can use a roulette wheel to select which NFTs can be integrated with each other. The fitness score of each NFT corresponds to a fusion probability: that is, the higher the fitness score, the higher the probability of being selected for fusion. Once the integration is complete, we get a new batch of NFTs and compete again. Each new competition in a genetic algorithm is called a generation. As generations pass, NFTs will evolve into more and more powerful NFTs.
from"fusion pool"The NFT generated in is different from the NFT initially dropped. These NFTs are generated via a genetic algorithm and inherit the skills of NFTs submitted to the pool. This ensures that these NFTs as rewards have a high economic value. In addition, the NFTs in the fusion pool are trained, while the NFTs in the initial airdrop are untrained.
One of the core functions of the fusion pool is to control the inflation of NFTs: if a player obtains a more powerful NFT in the fusion pool, the old NFT will be automatically destroyed
potential risks
The way AI learns is active learning, which requires a lot of human operations. Although the gameplay settings of AI Arena can stimulate users and increase their stay in the game, when acquiring customers in the early stage, it is necessary to effectively utilize the unique incentive mechanism of Web3 to quickly expand the player base and influence.
