Active Learing Decision Engines

A method for selecting a hypothesis from a set of hypotheses using an active learning decision engine, comprising:

• determining a plurality of equivalence classes using the active learning decision engine, where an equivalence class contains one or more hypotheses;
• determining a set of weighted edges using the active learning decision engine, where the weighted edges connect hypotheses in different equivalence classes and the weighted edges represent tests related to the connected hypotheses;
• removing weighted edges from the set of weighted edges which maximizes the computed expected reduction in weight using the active learning decision engine;
• observing the outcome of remaining tests associated with the weighted edges using the active learning decision engine;
• and
• selecting a hypothesis from the set of hypotheses based on the observed outcome using the active learning decision engine.

A method for selecting hypothesis from a set of hypotheses determined using observations of selected tests using an active learning decision engine comprising:

• selecting a hypothesis from the set of hypotheses using the active learning decision engine;
• for each test related to the selected hypothesis:
• selecting a test to run using the active learning decision engine, where the test is related to the selected hypothesis;
• running the selected test and observing the result of the selected test using the active learning decision engine;
• and
• computing the expected reduction in weight utilizing the observed result using the active learning decision engine;
• selecting the test which maximizes the computed expected reduction in weight using the active learning decision engine;
• adding the selected test to the answer using the active learning decision engine;
• observing the outcome of the selected test using the active learning decision engine;
• and
• selecting a hypothesis from the set of hypotheses using the observed outcome using the active learning decision engine.