Agent Demo

The following video from CSGO illustrates, the agent attempts to imitate the unique behavior of the human player based on his gameplay in CSGO setting: 15 hrs human player gameplay data and 30 hrs training time.

iAgent demo

Behaviors cloned:

• Obstacle avoidance

• Knee shots / Chest shots / Head shots

• Object Target Practice

Using behavioral cloning, a methodology developed by the iAgent protocol, neural networks and deep learning allows us to create agents that behave like gamers themselves, presenting gamers with challenging and interesting opponents.

The process of "behavioural cloning” involves teaching an intelligent model to mimic players' actions by observing examples. As a result, CSGO and CS2 agents are taught how to replicate human players' behavior, thought patterns and their actions.

The intelligent model can comprehend the underlying methods and patterns used by players through the analysis of an adequate amount of data.