This paper shows how agents' choice in communicative action can be designed to mitigate the effect of their resource 1/mits in the context of particular features of a collaborative planning task. I first motivate a number of hypotheses about effective language behavior based on a statistical analysis of a corpus of natural collaborative planning dialogues. These hypotheses are then tested in a dialogue testbed whose design is motivated by the corpus analysis. Experiments in the testbed examine the interaction between (1) agents' resource 1/mits in attentional capacity and inferential capacity; (2) agents' choice in communication; and (3) features of communicative tasks that affect task difficulty such as inferential complexity, degree of belief coordination required, and tolerance for errors. The results show that good algorithms for communication must be defined relative to the agents' resource 1/mits and the features of the task. Algorithms that are inefficient for inferentially simple, low coordination or fault-tolerant tasks are effective when tasks require coordination or complex inferences, or are fault-intolerant. The results provide an explanation for the occurrence of utterances in human dialogues that, prima facie, appear inefficient, and provide the basis for the design of effective algorithms for communicative choice for resource limited agents.

Centering was formulated as a model of the relationship between attentional state, the form of referring expressions, and the coherence of an utterance within a discourse segment (Grosz, Joshi and Weinstein, 1986; Grosz, Joshi and Weinstein, 1995). In this chapter, I argue that the restriction of centering to operating within a discourse segment should be abandoned in order to integrate centering with a model of global discourse structure. The within-segment restriction causes three problems. The first problem is that centers are often continued over discourse segment boundaries with pronominal referring expressions whose form is identical to those that occur within a discourse segment. The second problem is that recent work has shown that listeners perceive segment boundaries at various levels of granularity. If centering models a universal processing phenomenon, it is implausible that each listener is using a different centering algorithm.The third issue is that even for utterances within a discourse segment, there are strong contrasts between utterances whose adjacent utterance within a segment is hierarchically recent and those whose adjacent utterance within a segment is linearly recent. This chapter argues that these problems can be eliminated by replacing Grosz and Sidner's stack model of attentional state with an alternate model, the cache model. I show how the cache model is easily integrated with the centering algorithm, and provide several types of data from naturally occurring discourses that support the proposed integrated model. Future work should provide additional support for these claims with an examination of a larger corpus of naturally occurring discourses.

Human concept learning presents a version of the classic problem of induction, which is made particularly difficult by the combination of two requirements: the need to learn from a rich (i.e. nested and overlapping) vocabulary of possible concepts and the need to be able to generalize concepts reasonably from only a few positive examples. I begin this thesis by considering a simple number concept game as a concrete illustration of this ability. On this task, human learners can with reasonable confidence lock in on one out of a billion billion billion logically possible concepts, after seeing only four positive examples of the concept, and can generalize informatively after seeing just a single example. Neither of the two classic approaches to inductive inference -- hypothesis testing in a constrained space of possible rules and computing similarity to the observed examples -- can provide a complete picture of how people generalize concepts in even this simple setting. This thesis prop...

This paper develops a means-ends analysis of an inductive problem that arises in particle physics: how to infer from observed reactions conservation principles that govern all reactions among elementary particles. I show that there is a reliable inference procedure that is guaranteed to arrive at an empirically adequate set of conservation principles as more and more evidence is obtained. An interesting feature of reliable procedures for finding conservation principles is that in certain precisely defined circumstances they must introduce hidden particles. Among the reliable inductive methods there is a unique procedure that minimizes convergence time as well as the number of times that the method revises its conservation principles. Thus the aims of reliable, fast and steady convergence to an empirically adequate theory single out a unique optimal inference for a given set of observed reactions–including prescriptions for when exactly to introduce hidden particles.

. This paper examines psychological data on human reasoning with sets of negative defaults. A negative default is a statement of the form: Xs are typically not Ys. While there is pragmatic motivation for chaining positive defaults, chaining negative defaults #concluding from, As are typically not Bs and Bs are typically Cs that As are typically not Cs# is far less reasonable. Default inheritance reasoners universally prohibit `negativechaining'. However, examination of the psychological plausibility of various con#icting proof theories for default inheritance has demonstrated that some fundamental assumptions of the inheritance literature do not actually hold. This work has also revealed reasoning strategies which do describe human behaviors. In an e#ort to de#ne inheritance reasoners that are more predictiveofhuman reasoning with defaults, it is important to attend to these #ndings. This paper focuses on the fact that many people do in fact chain negative defaults. The pa...

Default inheritance reasoning is a propositional approach to nonmonotonic reasoning designed to model reasoning with natural language generics. Inheritance reasoners model sets of natural language generics as directed acyclic graphs, and inference corresponds to the specification of paths through those networks. A proliferation of inheritance proof theories exist in the literature along with extensive debate about the most reasonable way to construct inferences, based on intuitions about interpretations of particular inheritance networks. There has not been an accepted semantics for inheritance which unifies the set of possible proof theories, which would help identify truly ill-motivated proof theories. This thesis attempts to clarify the inheritance literature in the three ways indicated in the title: psychological plausibility, proof theory and semantics. The thesis intends to displace debate about the best inferences to draw about a network from logicians ' introspections to empiri...

Centering was formulated as a model of the relationship between attentional state, the form of referring expressions, and the coherence of an utterance within a discourse segment. In this paper, I argue that the restriction of centering to operating within a discourse segment should be abandoned in order to integrate centering with a model of global discourse structure. The within-segment restriction causes three problems. The first problem is that centers are often continued over discourse segment boundaries with pronominal referring expressions whose form is identical to those that occur within a discourse segment. The second problem is that recent work has shown that listeners perceive segment boundaries at various levels of granularity. If centering models human discourse processing, it is implausible that each listener is using a different centering algorithm.

In this paper we present a simple belief updating system using recurrent fuzzy rules which improves class prediction in ordered datasets. The recurrent fuzzy rule builds up belief in a class for each point in a sample-ordered or time-ordered dataset. Belief in each class is represented by a fuzzy set predicted class defined on the class universe. Belief in a class increases as positive cases are presented and decreases with negative cases. We show how the importance and characteristics of belief updating are determined through the generation of reference previous class fuzzy sets and evidential logic rule weights using mass assignment and semantic discrimination analysis. Performance of the recurrent belief updating rule is compared to the non-recurrent rule in application to facial feature detection and the classification of particles in gaseous streams. 1 Introduction Fuzzy rules which include a simple belief and memory term can perform markedly better at classification than fuzzy ...

Intuitions play a critical role in analytical philosophical activity. But do they qualify as genuine evidence for the sorts of conclusions philosophers seek? Skeptical arguments against intuitions are reviewed, and a variety of ways of trying to legitimate them are considered. A defense is off ered of their evidential status by showing how their evidential status can be embedded in a naturalistic framework.

My aim in this paper is to explore the relationship between epistemic rationality and instrumental rationality. By epistemic rationality, I mean, roughly, the kind of rationality which one displays when one believes propositions that are strongly supported by one's evidence and refrains from believing propositions that are improbable given one's evidence. Prominent