Abstract
This paper presents research results of a project that produced a progressively dependable knowledge amplifier. A knowledge amplifier can acquire a virtual-rule space automatically that is exponentially larger than the actual, declared-rule space and with a decreasing non-zero likelihood of error. A Knowledge Amplifier by Systematic Expert Randomization (KASER) overcomes the knowledge-acquisition bottleneck in intelligent systems by amplifying user-supplied knowledge. This enables the construction of an intelligent, creative system that fails softly, learns over a network, and has enormous potential for automated decision making. The system achieves creative reasoning through the use of declarative object trees. A KASER computes with words and phrases and presents a capability for analogical explanation. We implemented the KASER with an object-oriented design.