Qualitative simulation predicts the set of possible behaviors...

Learning, planning, and representing knowledge at multiple levels of temporal abstraction are key, longstanding challenges for AI. In this paper we consider how these challenges can be addressed within the mathematical framework of reinforcement learning and Markov decision processes (MDPs). We extend the usual notion of action in this framework to include options---closed-loop policies for taking action over a period of time. Examples of options include picking up an object, going to lunch, and traveling to a distant city, as well as primitive actions such as muscle twitches and joint torques. Overall, we show that options enable temporally abstract knowledge and action to be included in the reinforcement learning framework in a natural and general way. In particular, we show that options may be used interchangeably with primitive actions in planning methods such as dynamic programming and in learning methods such as Q-learning.

Geographers use spatial reasoning extensively in large-scale spaces, i.e., spaces that cannot be seen or understood from a single point of view. Spatial reasoning differentiates several spatial relations, e.g. topological or metric relations, and is typically formalized using a Cartesian coordinate system and vector algebra. This quantitative processing of information is clearly different from the ways humans draw conclusions about spatial relations. Formalized qualitative reasoning processes are shown to be a necessary part of Spatial Expert Systems and Geographic Information Systems. Addressing a subset of the total problem, namely reasoning with cardinal directions, a completely qualitative method, without recourse to analytical procedures, is introduced and a method for its formal comparison with quantitative formulae is defined. The focus is on the analysis of cardinal directions and their properties. An algebraic method is used to formalize the meaning of directions. The standard...

Many potential uses of qualitative physics, such as robot planning and intelligent computer-aided engineering, require integrating physics with actions taken by agents. This paper proposes to augment qualitative simulation to include the effects of actions to form action-augmented envisionments. The action-augmented envisionment incorporates both the effects of an agent's actions and what will happen in the physical world whether or not the agent does something. Consequently, it should provide a richer basis for planning and procedure generation than any previous representation. This paper defines actionaugmented envisionments and an algorithm for directly computing them, along with an analysis of its complexity and suitability for different kinds of problems. We describe our initial implementation and discuss potential extensions, including incremental algorithms. Keywords: Qualitative reasoning, planning, artificial intelligence. Presented at the 2nd Qualitative Physics Workshop Pa...

Incomplete knowledge of the structure of mechanisms is an important fact of life in reasoning, commonsense or expert, about the physical world. Qualitative simulation captures an important kind of incomplete, ordinal, knowledge, and predicts the set of qualitatively possible behaviors of a mechanism, given a qualitative description of its structure and initial state. However, one frequently has quantitative knowledge as well as qualitative, though seldom enough to specify a numerical simulation.

A central goal of qualitative physics is to provide a framework for organizing and using quantitative knowledge. One important use of quantitative knowledge is numerical simulation. While current numerical simulators are powerful, they are often hard to con struct, do not reveal the assumptions underlying their construction, and do not produce explanations of the behaviors they predict. This paper shows how to combine qualitative and quantitative models to produce a new class of self-explanatory simulations which combine the advantages of both kinds of reasoning. Self-explanatory simulations provide the accuracy of numerical models and the interpretive power of qualitative reasoning. We define what self-explanatory simulations are and show how to construct them automatically. We illustrate their power with some examples generated with an implemented system, SIMGEN. We analyze the limitations of our techniques, and discuss plans for future work.

Reaction-time and eye-fixation data are analyzed to investigate how people infer the kinematics of simple mechanical systems (pulley systems) from diagrams showing their static configuration. It is proposed that this mental animation process involves decomposing the representation of a pulley system into smaller units corresponding to the machine components and animating these components in a sequence corresponding to the causal sequence of events in the machine's operation. Although it is possible for people to make inferences against the chain of causality in the machine, these inferences are more difficult, and people have a preference for inferences in the direction of causality. The mental animation process reflects both capacity limitations and limitations of mechanical knowledge. Understanding the operation of deterministic systems, such as mechanical or electronic devices, includes the ability to infer the state of one component of the system given information about the states of the other system components and the relations between the components. This type of understanding is central to how people design, troubleshoot, and operate devices. This article describes how people infer the motion of components of a simple mechanical system (a pulley system) from knowledge of the configuration of the system and the movement of one of the system components. It provides an account of the process of inferring motion, the type of knowledge that allows people to infer motion, and the characteristics of human information processing that constrain the inference process. I refer to this process as mental animation.

Qualitative models arising in the artificial intelligence domain often concern real systems that are difficult to represent with traditional means. However, some promise for dealing with such systems is offered by research in simulation methodology. Such research produces models that combine both continuous and discrete event formalisms. Nevertheless, the aims and approaches of the AI and the simulation communities remain rather mutually ill-understood. Consequently, there is a need to bridge theory and methodology in order to have a uniform language when either analyzing or reasoning about physical systems. This article introduces a methodology and formalism for developing multiple, cooperative models of physical systems of the type studied in qualitative physics. The formalism combines discrete event and continuous models and offers an approach to building intelligent machines capable of physical modeling and reasoning. Categories and Subject Descriptors: I.2.4 [Artificial Intelligen...

Qualitative reasoning can, and should, be decomposed into a model-building task, which creates a qualitative differential equation (QDE) as a model of a physical situation, and a qualitative simulation task, which starts with a QDE, and predicts the possible behaviors following from the model. In support of this claim, we present QPC, a model builder that takes the general approach of Qualitative Process Theory [ Forbus, 1984 ] , describing a scenario in terms of views, processes, and influences. However, QPC builds QDEs for simulation by QSIM, which gives it access to a variety of mathematical advances in qualitative simulation incorporated in QSIM. We present QPC and its approach to Qualitative Process Theory, provide an example of building and simulating a model of a non-trivial mechanism, and compare the representation and implementation decisions underlying QPC with those of QPE [ Falkenhainer and Forbus, 1988; Forbus, 1990 ] . Introduction There have been a variety of producti...

Abstract not found