1.WHAT RESEARCHERS DO FOR AI AGENTS?
Researchers have developed a new approach to give AI agents a prescient perspective. Their machine learning framework allows cooperative or competitive AI agents to think about what other agents will do when time approaches infinity rather than the next few steps. Agents then adjust their behavior accordingly, influencing the future behavior of other agents to arrive at the optimal long-term solution.

2.WHICH WAY AI AGENT LEARNING MULTI AGENT REINFORCEMENT?
Researcher focusing on multi-agent reinforcement learning. Reinforcement learning is a form of machine learning in which AI agents learn through trial and error. Researchers reward agents for “good” behavior that helps them achieve their goals. Agents adjust their behavior to maximize this reward and eventually become task experts.

They developed a machine learning framework known as FURTHER that allows agents to learn how to coordinate their behavior as they interact with other agents to achieve this active equilibrium. FURTHER uses two machine learning modules for this. The first inference engine allows agents to infer the future behavior of other agents and the learning algorithms they use based solely on their past behavior.

3.HOW THE FARSIGHTED AI AGENTS WORK?
The framework allows groups of autonomous drones to work together to locate a hiker lost in a dense forest, or search for other vehicles that may be traveling on a busy highway. It can be used in self-driving cars that try to keep passengers safe by predicting future movements.