Creativity is sometimes taken to be an inexplicable aspect of human activity. By summarizing a considerable body of literature on creativity, I hope to show how to turn some of the best ideas about creativity into programs that are demonstrably more creative than any we have seen to date. I believe the key to building more creative programs is to give them the ability to reflect on and modify their own frameworks and criteria. That is, I believe that the key to creativity is at the metalevel.
There is general agreement that knowledge plays a key role in intelligent behavior, but most work on this topic has emphasized domain-specific expertise. We argue, in contrast, that cognitive systems also benefit from meta-level knowledge that has a domain-independent character. In this paper, we propose a representational framework that distinguishes between these two forms of content, along with an integrated architecture that supports their use for abductive interpretation and hierarchical skill execution. We demonstrate this framework's viability on high-level aspects of extended dialogue that require reasoning about, and altering, participants' beliefs and goals. Furthermore, we demonstrate its generality by showing that the meta-level knowledge operates with different domain-level content. We conclude by reviewing related work on these topics and discussing promising directions for future research.
Rippling is a radically new technique for the automation of mathematical reasoning. It is widely applicable whenever a goal is to be proved from one or more syntactically similar givens. It was originally developed for inductive proofs, where the goal was the induction conclusion and the givens were the induction hypotheses. It has proved to be applicable to a much wider class of tasks, from summing series via analysis to general equational reasoning. The application to induction has especially important practical implications in the building of dependable IT systems, and provides solutions to issues such as the problem of combinatorial explosion.
Sequential decision problems are often approximately solvable by simulating possible future action sequences. Metalevel decision procedures have been developed for selecting which action sequences to simulate, based on estimating the expected improvement in decision quality that would result from any particular simulation; an example is the recent work on using bandit algorithms to control Monte Carlo tree search in the game of Go. In this paper we develop a theoretical basis for metalevel decisions in the statistical framework of Bayesian selection problems, arguing (as others have done) that this is more appropriate than the bandit framework. We derive a number of basic results applicable to Monte Carlo selection problems, including the first finite sampling bounds for optimal policies in certain cases; we also provide a simple counterexample to the intuitive conjecture that an optimal policy will necessarily reach a decision in all cases. We then derive heuristic approximations in both Bayesian and distribution-free settings and demonstrate their superiority to bandit-based heuristics in one-shot decision problems and in Go.
Embedded systems consisting of collaborating agents capable of interacting with their environment are becoming ubiquitous. It is crucial for these systems to be able to adapt to the dynamic and uncertain characteristics of an open environment. In this paper, we argue that multiagent meta-level control (MMLC) is an effective way to determine when this adaptation process should be done and how much effort should be invested in adaptation as opposed to continuing with the current action plan. We describe a reinforcement learning based approach to learn decentralized meta-control policies offline. We then propose to use the learned reward model as input to a global optimization algorithm to avoid conflicting meta-level decisions between coordinating agents. Our initial experiments in the context of NetRads, a multiagent tornado tracking application show that MMLC significantly improves performance in a 3-agent network.