REAL TIME CONTEXT LEARNING BY SOFTWARE AGENTS

摘要

Providing dynamic learning for software agents (140) in a simulation. Software agents (140) with learners (165) are capable of learning from examples. When a non-player character (142) queries the learner (165), it can provide a next action similar to the player character (141). The game designer provides program code, from which compile-time steps determine a set of raw features (150). The code might identify a function (like computing distances). At compile-time steps, determining these raw features (150) in response to a scripting language, so the designer can specify which code should be referenced. A set of derived features (160), responsive to the raw features (150), might be relatively simple, more complex, or determined in response to a learner (165). The set of such raw (150) and derived features (160) form a context for a learner (165). Learners (165) might be responsive to (more basic) learners, to results of state machines, to calculated derived features (160), or to raw features (150). The learner (165) includes a machine learning technique.