Lenhart K. Schubert

The TRAINS Project: A Case Study in Defining a Conversational Planning Agent

The TRAINS project is an effort to build a conversationally proficient planning assistant. A key part of the project is the construction of the TRAINS system, which provides the research platform for a wide range of issues in natural language understanding, mixed-initiative planning systems, and representing and reasoning about time, actions and events. Four years have now passed since the beginning of the project. Each year we have produced a demonstration system that focused on a dialog that illustrates particular aspects of our research. The commitment to building complete integrated systems is a significant overhead on the research, but we feel it is essential to guarantee that the results constitute real progress in the field. This paper describes the goals of the project, and our experience with the effort so far. .pp This paper is to appear in the Journal of Experimental and Theoretical AI, 1995.

Extending the expressive power of semantic networks

Abstract “Factual knowledge” used by natural language processing systems can be conveniently represented in the form of semantic networks. Compared to a “linear” representation such as that of the Predicate Calculus however, semantic networks present special problems with respect to the use of logical connectives, quantifiers, descriptions, and certain other constructions. Systematic solutions to these problems will be proposed, in the form of extensions to a more or less conventional network notation. Predicate Calculus translations of network propositions will frequently be given for comparison, to illustrate the close kinship of the two forms of representation.

Generically Speaking, or, Using Discourse Representation Theory to Interpret Generics

In this paper we discuss a proposal for representing the logical form of generic and habitual sentences. The proposal is a development of Schubert and Pelletier (1987), although it stands on its own and can be comprehended without having first read the earlier paper. (Most of what would be missed would merely be a discussion of background literature, and reasons to want a different logical form from those extant in the literature).

Efficient algorithms for qualitative reasoning about time

Reasoning about temporal information is an important task in many areas of Artificial Intelligence. In this paper we address the problem of scalability in temporal reasoning by providing a collection of new algorithms for efficiently managing large sets of qualitative temporal relations. We focus on the class of relations forming the Point Algebra (PA-relations) and on a major extension to include binary disjunctions of PA-relations (PA-disjunctions). Such disjunctions add a great deal of expressive power, including the ability to stipulate disjointness of temporal intervals, which is important in planning applications. .pp Our representation of time is based on timegraphs, graphs partitioned into a set of chains on which the search is supported by a metagraph data structure. The approach is an extension of the time representation proposed by Schubert, Taugher and Miller in the context of story comprehension. The algorithms herein enable construction of a timegraph from a given set of PA-relations, querying a timegraph, and efficiently checking the consistency of a timegraph augmented by a set of PA-disjunctions. Experimental results illustrate the efficiency of the proposed approach.

An optimal algorithm for constructing the Delaunay triangulation of a set of line segments

In this paper, we first define a new Voronoi diagram for the endpoints of a set of line segments in the plane which do not intersect (except possibly at their endpoints), which is called a <italic>bounded Voronoi diagram</italic>. In this Voronoi diagram, the line segments themselves are regarded as obstacles. We present an optimal &THgr;(<italic>n log n</italic>) algorithm to construct it, where <italic>n</italic> is the number of input line segments. We then show that the straight-line dual of the bounded Voronoi diagram of a set of non-intersecting line segments is the Delaunay triangulation of that set. And the straight-line dual can be obtained in time proportional to the number of input line segments when the corresponding bounded Voronoi diagram is available. Consequently, we obtain an optimal &THgr;(<italic>n log n</italic>) algorithm to construct the Delaunay triangulation of a set of <italic>n</italic> non-intersecting line segments in the plane. Our algorithm improves the time bound <italic>&Ogr;</italic>(<italic>n</italic><supscrpt>2</supscrpt>) of the previous best algorithm.

THE STRUCTURE AND ORGANIZATION OF A SEMANTIC NET FOR COMPREHENSION AND INFERENCE

We have developed a network representation for propositional knowledge that we believe to be capable of encoding any proposition expressible in natural language. The representation can be regarded as a computer-oriented logic with associative access paths from concepts to propositions. Its syntax is closely modeled on predicate calculus but includes constructs for expressing some kinds of vague and uncertain knowledge. The representation allows the encoding and efficient use of caselike semantic constraints on predicate arguments for the purpose of language comprehension: these constraints are simply implications of the predicates concerned. Our approach to language comprehension is based on nonprimitive representations. We argue that primitive representations of simple propositions are often extremely complex, and offer no real advantages. We have demonstrated these ideas with a mini-implementation capable of mapping certain kinds of declarative sentences into the network representation. The implementation emphasizes the proper handling of iterated adjectival modifiers, especially comparative modifiers. More recently, we have worked on the problem of rapid access to the facts that are relevant to a query. Our solution involves the use of back-link structures from concepts to propositions, called “topic access skeletons,” which conform with general topic hierarchies in memory. For example, the proposition “Clyde is grey” is classified under the “coloring” topic for Clyde, which is subsumed under the “appearance” topic, and in turn under the “external quality” topic, and finally under the “physical quality” topic for Clyde. The form of a query (or of an assertion) can be used to determine what concepts in memory should be accessed as starting points, and what paths in the associated access skeletons should be followed in order to access the relevant information. We have demonstrated the feasibility of building such hierarchies, inserting information into them automatically, and accessing the inserted information with a second experimental implementation. The hierarchic organization appears capable of providing order-of-magnitude improvements in question-answering efficiency, with only a doubling in storage costs.

Time revisited

Temporal reasoning is essential for many artificial intelligence applications. To date, most research has concentrated on temporal inference in isolation without considering the role it can play in a more general reasoning environment. This paper takes an efficient temporal reasoner and extends its inferential capabilities to handle both strict and nonstrict relations. The resulting temporal specialist is incorporated into a system intended for low‐level reasoning in natural language understanding. The specialist assists the resolution‐based theorem prover in function evaluation, literal evaluation, and generalized resolving and factoring. The combined system can do some proofs in just a few steps that would normally require many. An example from the fully operational hybrid system is included.

TENSE TREES AS THE "FINE STRUCTURE" OF DISCOURSE

We present a new compositional tense-aspect deindexing mechanism that makes use of tense trees as components of discourse contexts. The mechanism allows reference episodes to be correctly identified even for embedded clauses and for discourse that involves shifts in temporal perspective, and permits deindexed logical forms to be automatically computed with a small number of deindexing rules.

Monotonic Solution of The Frame Problem in The Situation Calculus

The frame problem originally surfaced within McCarthy’s Situation Calculus [McCarthy68], when [McCarthy&Hayes69] applied it to reasoning about goal achievement. To illustrate their approach, they considered the problem of initiating a telephone conversation. They began by writing down plausible axioms which seemed to characterize the preconditions and effects of looking up a person’s telephone number and dialling that number. However, they found that they were still unable to prove that the plan “look up the number and dial it” would work, even if all the initial conditions were right (i.e., that the caller had a telephone and a telephone book, that the intended party was home, etc.). For example, the axioms provided no assurance that looking up the number would not make the caller’s telephone disappear, thus voiding a precondition for dialling.

Interpreting Tense, Aspect and Time Adverbials: A Compositional, Unified Approach

We extend our theory of English tense, aspect and time adverbials [Hwang and Schubert, 1992, 1993] to deal with a wider range of time adverbials, including many adverbials of frequency, cardinality, duration, and time span, and adverbials of temporal relation involving subordinating conjunctions such as after, since, and until. Our theory is fully formal in that it derives indexical (quasi-)logical forms from syntactic-semantic rule pairs of a formal grammar, and nonindexical logical forms via deindexing rules in the form of equivalences and equations. The grammar allows for complex sentences and the semantic rules and deindexing rules are easy to implement computationally, producing formulas in Episodic Logic.