Property:Abstract
From Tetherless World Wiki
Facts about AbstractRDF feed
| Has typeThis property is a special property in this wiki. | TextThis type is among the standard datatypes of this wiki. |
(previous 25) (next 25)
Pages using the property "Abstract"
Showing 25 pages using this property.
" | |
|---|---|
| "Reducing" CLASSIC to Practice: Knowledge Representation Theory Meets Reality + | Most recent key developments in research o … Most recent key developments in research on knowledge representation (KR) have been of the more theoretical sort, involving worst-case complexity results, solutions to technical challenge problems, etc. While some of this work has influenced practice in Artificial Intelligence, it is rarely if ever made clear what is compromised when the transition is made from relatively abstract theory to the real world. CLASSIC is a description logic with an ancestry of extensive theoretical work (tracing back over twenty years to KL-ONE), and several novel contributions to KR theory. Basic research on CLASSIC paved the way for an implementation that has been used significantly in practice, including by users not versed in KR theory. In moving from a pure logic to a practical tool, many compromises and changes of perspective were necessary. We report on this transition and articulate some of the profound influences practice can have on relatively idealistic theoretical work. We have found that CLASSIC has been quite useful in practice, yet still strongly retains most of its original spirit, but much of our thinking and many details had to change along the way. many details had to change along the way. |
A | |
| A Bayesian Analysis of Randomized Approximation Procedures + | Computer scientists have used Zero-One Est … Computer scientists have used Zero-One Estimation Theory to develop a body of results on the convergence of Monte-Carlo algorithms. In much of this work, investigators have analyzed bounds on the variance of an estimator p after N trials assuming knowledge of the expectation u of a random variable. We present an alternative framework for analyzing randomized approximation procedures that considers the manner in which u are distributed for Bernoulli processes. Rather than relying on results about error that follow from the assumption of independent and identically distributed random variables, we consider the properties of the beta distribution. The beta is a conjugate distribution for the Bernoulli sampling processes. Conjugate distributions allow us to represent both the prior and posterior probability distributions for a sampling process. We derive probabilistic stopping rules for Monte-Carlo algorithms by approximating the portion of the cumulative distribution of the beta defined by bounds on error. The probabilistic stopping rules do not require prior knowledge or estimates of u. We prove an upper bound on N that is a function of the value of the estimator p rather than u. For small relative error or u, our stopping rule gives an O(log n) speed-up on the running time of Monte-Carlo algorithms over the running time predicted by the Zero-One Estimator Theorem. edicted by the Zero-One Estimator Theorem. |
| A Bayesian Analysis of Simulation Algorithms for Inference in Belief Networks + | Belief networks are an expressive represen … Belief networks are an expressive representation for encoding expert knowledge about uncertain causal relationships. Both exact and approximation methods for performing inference with belief networks can pose difficult computational problems in the worst case. Nevertheless, approximation procedures hold promise to provide estimates efficiently for a variety of complex networks that resist exact solution. We characterize the performance of algorithms in the important class of inference procedures based on stochastic simulation. We develop terms for the error associated with estimates generated by several simulation methods, including forward simulation, likelihood weighting, and randomized approximation strategies. , and randomized approximation strategies. |
| A Bayesian Computer-Based Approach to the Physician's Use of the Clinical Research Literature + | No abstract available |
| A Bayesian Method for the Induction of Probabilistic Networks from Data + | This paper presents a Bayesian method for … This paper presents a Bayesian method for constructing probabilistic networks from a database of cases. In particular, we focus on constructing Bayesian belief networks. Applications include computer-assisted hypothesis testing, automated scientific discovery, and automated construction of probabilistic expert systems. We extend the basic method to handle missing data and hidden (latent) variables. We show how to perform probabilistic inference by averaging over the inferences of multiple belief networks. Results are presented of a preliminary evaluation of an algorithm for constructing a belief network from a database of cases. Finally, we relate the methods in this paper to previous work, and we discuss open problems. evious work, and we discuss open problems. |
| A Bayesian Perspective in Confidence + | We present a representation of partial con … We present a representation of partial confidence in belief and preference that is consistent with the tenets of decision-theory. The fundamental insight underlying the representation is that if a person is not completely confident in a probability or utility assessment, additional modeling of the assessment may improve decisions to which it is relevant. We show how a traditional decision-analytic approach can be used to balance the benefits of additional modeling with associated costs. The approach can be used during knowledge acquisition to focus the attention of a knowledge engineer or expert on parts of a decision model that deserve additional refinement. model that deserve additional refinement. |
| A Belief Network Model for Interpretation of ICU Data + | Belief networks provide a causal probabili … Belief networks provide a causal probabilistic framework for the representation of medical knowledge. We have developed VPnet, a belief-network model of the pathophysiology of patients in the intensive-care unit (ICU), and have incorporated this belief-network in a system _VentPlan_ that assists in the care and monitoring of patients in the ICU. VPnet converts patient observations into probability distributions for a set of physiological parameters used by VentPlan's mathematical model. VPnet represents the uncertainty of data observations explicitly and implements a model of increasing uncertainty as the time from an observation increases. We have evaluated VPnet using sets of inputs corresponding to a variety of clinical states, and we show calculated physiologic parameter distributions appropriate for the clinical state. Evaluation of complex belief-network models is difficult due to the lack of a gold standard for comparison, and because there is a large number of possible sets of input states. e number of possible sets of input states. |
| A Brief Guide to MAITRE and MODEL: An Ontology Editor and a Frame-Based Knowledge Representation Language + | This document provides a brief introductio … This document provides a brief introduction to the use of the MAITRE tool. MAITRE is an ontology editor for use within the PROTEGE-II architecture; it allows the user to model knowledge in a domain with a hierarchy of classes. This system uses a frame-based knowledge-representation language called MODEL. The goal of MAITRE is to allow the user to edit and build MODEL code that can then be used to build domain-specific knowledge-acquisition tools. main-specific knowledge-acquisition tools. |
| A Categorization of Explanation Questions for Task Processing Systems + | A critical aspect of any explanation modul … A critical aspect of any explanation module is the set of user questions the system will be able to address. However, there has been relatively little work on listing and organizing the various categories of questions helpful to explanation. In this paper we address this problem by proposing a categorization of question types relevant to explaining task processing. For each question type, we also propose alternative explanation strategies for answering them. This categorization has helped to drive our work on an integrated cognitive explanation environment that has been used to explain the behavior of CALO, a software cognitive assistant that learns and organizes. itive assistant that learns and organizes. |
| A Causal Functional Representation Language with Behavior-Based Semantics + | Understanding the design of a device requi … Understanding the design of a device requires both knowledge of the general physical principles that determine its behavior and knowledge of its intended functions. However, the majority of work in model-based reasoning has focused on using either one of these types of knowledge alone. In order to use both types of knowledge in understanding a device design, one must represent the functional knowledge in such a way that it has a clear interpretation in terms of observed behavior. We propose a new formalism, Causal Functional Representation Language (CFRL), for representing device functions with well-defined semantics in terms of behavior. CFRL allows the specification of conditions that a behavior must satisfy, such as occurrence of temporal sequences of events and causal relations among them and the components. We have used CFRL as the basis for a functional verification program, which determines whether a behavior achieves an intended function. a behavior achieves an intended function. |
| A Combination of Cutset Conditioning with Clique-Tree Propagation in the Pathfinder System + | Cutset conditioning and clique-tree propag … Cutset conditioning and clique-tree propagation are two popular methods for performing exact probabilistic inference in Bayesian belief networks. Cutset conditioning is based on decomposition of a subset of network nodes,whereas clique-tree propagation depends on aggregation of nodes. We describe a means to combine cutset conditioning and clique-tree propagation in an approach called AGGREGATION AFTER DECOMPOSITION (AD). We discuss the application of the AD method in the Pathfinder system, a medical expert system that offers assistance with diagnosis in hematopathology. istance with diagnosis in hematopathology. |
| A Comparison of Action-Based Hierarchies and Decision Trees for Real-Time Performance + | Decision trees have provided a classical m … Decision trees have provided a classical mechanism for progressively narrowing down a search from a large group of possibilities to a single alternative. The structuring of a decision tree is based on a heuristic that maximizes the value of the information gained at each level in the hierarchy. Decision trees are effective when an agent needs to reach the goal of complete diagnosis as quickly as possible and cannot accept a partial solution. We present an alternative to the decision tree heuristic which is useful when partial solutions do have value and when limited resources may require an agent to accept a partial solution. Our heuristic maximizes the improvement in the value of the partial solution gained at each level in the hierarchy; we term the resulting structure an action-based hierarchy. We present the results of a set of experiments designed to compare these two heuristics for hierarchy structuring. Finally, we describe some preliminary work we have done in applying these ideas to a medical domain--surgical intensive care unit (SICU) patient monitoring. nsive care unit (SICU) patient monitoring. |
| A Comparison of Anapron with Seven Other Name-Pronounciation Systems + | Anapron is a name-pronunciation system bas … Anapron is a name-pronunciation system based on a general method for combining rule-based and case-based reasoning. An experiment was run to see how this system compares with existing systems for name pronunciation. Seven such systems were tested: three state-of-the-art commercial systems (from Bellcore, Bell Labs, and DEC), two variants of a machine-learning system(NETtalk), and two humans. Each system was run on the same 400-name test set.The acceptability of its pronunciations was evaluated by a panel of 14 test subjects. To hide the identities of the systems, the order of systems was randomized for each test name, and all pronunciations were read by the DECtalk speech synthesizer. The main result was that Anapron was found to perform almost at the level of the commercial systems, and significantly better than the two versions of NETtalk. y better than the two versions of NETtalk. |
| A Comparison of Two Computer-Based Prognostic Systems for AIDS + | We compare the performances of a Cox model … We compare the performances of a Cox model and a neural network model that are used as prognostic tools for a cohort of people living with AIDS. We modeled disease progression for patients who had AIDS (according to the 1993 CDC definition) in a cohort of 588 patients in California, using data from the ATHOS project. We divided the study population into 10 training and 10 test sets and evaluated the prognostic accuracy of a Cox proportional hazards model and of a neural network model by determining the number of predicted deaths, the sensitivities, specificities, positive predictive values, and negative predictive values for intervals of one year following the diagnosis of AIDS. For the Cox model, we further tested the agreement between a series of binary observations, representing death in one, two, and three years, and a set of estimates which define the probability of survival for those intervals. Both models were able to provide accurate numbers on how many patients were likely to die at each interval, and reasonable individualized estimates for the two- and three-year survival of a given patient, but failed to provide reliable predictions for the first year after diagnosis. There was no evidence that the Cox model performed better than did the neural network model or vice-versa, but the former method had the advantage of providing some insight on which variables were most influential for prognosis. Nevertheless, it is likely that the assumptions required by the Cox model may not be satisfied in all data sets, justifying the use of neural networks in certain cases. e use of neural networks in certain cases. |
| A Comparison of the Temporal Expressiveness of Three Database Query Methods + | Time is a multifaceted phenomenon that dev … Time is a multifaceted phenomenon that developers of clinical decision-support systems can model at various levels of complexity. An unresolved issue for the design of clinical databases is whether the underlying data model should support interval semantics. In this paper, we examine whether interval-based operations are required for querying protocol-based conditions. We report on an analysis of a set of 256 eligibility criteria that the T-HELPER system uses to screen patients for enrollment in eight clinical-trial protocols for HIV disease. We consider three data-manipulation methods for temporal querying: the consensus query representation Arden Syntax, the commercial standard query language SQL, and the temporal query language TimeLineSQL (TLSQL). We compare the ability of these three query methods to express the eligibility criteria. Seventy nine percent of the 256 criteria require operations on time stamps. These temporal conditions comprise four distinct patterns, two of which use interval-based data. Our analysis indicates that the Arden Syntax can query the two non-interval patterns, which represent 54% of the temporal conditions. Timepoint comparisons formulated in SQL can instantiate the two non-interval patterns and one interval pattern, which encompass 96% of the temporal conditions. TLSQL, which supports an interval-based model of time, can express all four types of temporal patterns. Our results demonstrate that the T-HELPER system requires simple temporal operations for most protocol-based queries. Of the three approaches tested, TLSQL is the only query method that is sufficiently expressive for the temporal conditions in this system. or the temporal conditions in this system. |
| A Component-Based Approach to Automation of Protocol-Directed Therapy + | Objective: Automating the task of plannin … Objective: Automating the task of planning protocol-directed therapy requires a computer program to take as input clinical data stored in an electronic patient-record system, and to generate as output recommendations for therapeutic interventions and laboratory testing that are defined by predefined protocols. The output must be tailored for the current patient situation and stage of protocol execution. Our goal has been to model the functional requirements of the therapy-planning task.Design: We constructed a computational model that includes components that(1) interpret abstract protocol specifications to construct appropriate patient-specific treatment plans, (2) infer from time-stamped patient data higher-level, interval-based, abstract concepts, (3) perform time-oriented queries on a time-oriented patient database, and (4) allow for acquisition and maintenance of protocol knowledge in a manner that facilitates efficient processing both by humans and computers. We have implemented these components in a computer system known as EON.Results: The EON architecture brings together (1) a therapy planner based on a reusable problem-solving method known as episodic skeletal-plan refinement,(2) the RESUME temporal-abstraction system, which implements the knowledge-based temporal-abstraction problem-solving method, (3) the Chronus database system, which processes complex temporal queries referred to a clinical, time-oriented relational database, and (4) special-purpose knowledge-acquisition tools that are generated automatically from descriptions of the relevant clinical domains by the PROTEGE-II system. We have evaluated the capabilities of the EON components by implementing T-Helper, a computer-based patient record system that uses EON to offer advice regarding the management of patients who have AIDS and HIV infection.Conclusion: Each of the modules that comprise the EON architecture has been developed as a self-contained, reusable component. The integration of these elements, however, leads to an aggregate component that is itself reusable for automating the task of planning protocol-based therapy in a variety of clinical domains. EON provides a comprehensive, yet flexible approach that facilitates development, maintenance, and execution of electronic knowledge bases that encode clinical protocols of substantial complexity.Objective: Automating the task of planning protocol-directed therapy requires a computer program to take as input clinical data stored in an electronic patient-record system, and to generate as output recommendations for therapeutic interventions and laboratory testing that are defined by predefined protocols. The output must be tailored for the current patient situation and stage of protocol execution. Our goal has been to model the functional requirements of the therapy-planning task.Design: We constructed a computational model that includes components that(1) interpret abstract protocol specifications to construct appropriate patient-specific treatment plans, (2) infer from time-stamped patient data higher-level, interval-based, abstract concepts, (3) perform time-oriented queries on a time-oriented patient database, and (4) allow for acquisition and maintenance of protocol knowledge in a manner that facilitates efficient processing both by humans and computers. We have implemented these components in a computer system known as EON.Results: The EON architecture brings together (1) a therapy planner based on a reusable problem-solving method known as episodic skeletal-plan refinement,(2) the RESUME temporal-abstraction system, which implements the knowledge-based temporal-abstraction problem-solving method, (3) the Chronus database system, which processes complex temporal queries referred to a clinical, time-oriented relational database, and (4) special-purpose knowledge-acquisition tools that are generated automatically from descriptions of the relevant clinical domains by the PROTEGE-II system. We have evaluated the capabilities of the EON components by implementing T-Helper, a computer-based patient record system that uses EON to offer advice regarding the management of patients who have AIDS and HIV infection.Conclusion: Each of the modules that comprise the EON architecture has been developed as a self-contained, reusable component. The integration of these elements, however, leads to an aggregate component that is itself reusable for automating the task of planning protocol-based therapy in a variety of clinical domains. EON provides a comprehensive, yet flexible approach that facilitates development, maintenance, and execution of electronic knowledge bases that encode clinical protocols of substantial complexity. nical protocols of substantial complexity. |
| A Component-Based Architecture for Automation of Protocol-Directed Therapy + | The automation of protocol-based care requ … The automation of protocol-based care requires reasoning about a patient's situation over time and about how the standard protocol plan can be adapted to address the patient's current clinical situation. The EON architecture brings together (1) a skeletal-planning reasoning method, ESPR, that can determine appropriate clinical interventions by instantiating an abstract protocol specification, (2) a temporal-reasoning system, RESUME, that can infer from time-stamped patient data higher-level, interval-based concepts, and (3) a historical database system, Chronus, that can perform temporal queries on a database of interval-based patient descriptions. The modular problem-solving elements of EON operate on knowledge bases of clinical protocols that clinicians enter into domain-specific knowledge-acquisition tools generated by the PROTEGE-II system. The EON architecture provides an integrated framework for development, execution, and maintenance of clinical-protocol knowledge bases. ance of clinical-protocol knowledge bases. |
| A Compositional Approach to Causality + | Inferring causality from equation models c … Inferring causality from equation models characterizing engineering domains is important towards predicting and diagnosing system behavior. Most previous attempts in this direction have failed to recognize the key differences between equations which model physical phenomena and those that just express rationality or numerical conveniences of the designer. These different types of equations bear different causal implications among the model parameters they relate. We show how unstructured and ad hoc formulations of equation models for apparent numerical conveniences are lossy in the causal information encoding and justify the use of CML as a model formulation paradigm which retains these causal structures among model parameters by clearly separating equations corresponding to phenomena and rationality. We provide an algorithm to infer causality from the active model fragments by using the notion of PreCondition graphs. y using the notion of PreCondition graphs. |
| A Comprehensive Methodology for Building Hybrid Models of Physcial Systems + | This paper describes a comprehensive model … This paper describes a comprehensive model building framework for mixed continuous/discrete, i.e., hybrid physical system models. Hybrid models arise naturally when modeling embedded systems (physical systems with digital controllers) and complex physical systems whose behavior may be simplified by introducing discrete transitions to replace fast nonlinear dynamics. We study time scale and parameter abstraction methods, and derive transition semantics that are consistent with physical system principles. These are translated into a systematic modeling methodology with formal execution semantics for deriving hybrid behaviors in three distinct modes of operation: (i) continuous, (ii) pinnacles, and (iii) mythical. Behavior in the continuous modes is governed by a system of differential algebraic equations (DAEs). Pinnacles, an artifact of time scale abstraction, compress behaviors over small intervals to a point in real time, and define a discrete switching model based on a priori state values. Mythical modes, an artifact of parameter abstractions, involve transitions through modes that have no real existence on the time line. Discrete switching transitions for mythical modes are defined in terms of a posteriori state values. In conjunction with the formal mathematical specification language for hybrid models, we also derive a set of model verification procedures, the principles of invariance of state, divergence of time, and temporal evolution of state, that provide the framework for designing hybrid simulators. We adopt hybrid automata as the computational model for our hybrid specification language. We have applied this model building framework in a number of different domains. Recent work has focused on developing a compositional modeling framework to facilitate the model construction task. to facilitate the model construction task. |
| A Computer Program for Statistically Based Decision Analysis + | The majority of patients with coronary art … The majority of patients with coronary artery disease do not fall into the well defined populations from randomized clinical trials. Observational databases contain a rich source of information that could be used by practicing physicians to evaluate treatment alternatives for their patients. We describe a computer system, the CABG Kibitzer, which uses an integrated approach to evaluate the treatment alternatives for CAD patients. We combine a statistical multivariate model for calculating survival advantages with DA techniques for assessing patient preferences and sensitivity analysis, to create one tool that physicians find easy to use in daily clinical practice. The development of tools of this kind is a necessary step in making the data of outcome studies accessible to practicing physicians. udies accessible to practicing physicians. |
| A Computer-Based Interview to Identify HIV Risk Behaviors and to Assess Patient Preferences for HIV-Related Health States + | We developed a computer-based utility asse … We developed a computer-based utility assessment tool to assess the preferences of patients towards HIV -related health states and identify risk behaviors (both sexual and drug related) of the patient being interviewed. The reliability of the computer-based interview was assessed through comparison with person-to-person interviews.Our pilot study included 22 patients. Twelve of these patients were also interviewed by the research assistants in person-to-person interviews. The agreement between the person-to-person and computer-based interviews was excellent (3 discrepancies of 180 compared answers), and the majority of the patients preferred to use the computer to disclose sensitive information regarding risk behaviors. Our study suggests that assessment of patient preferences and risk factors can be performed reliably through a computer-based interview. liably through a computer-based interview. |
| A Computer-Based Tool for Generation of Progress Notes + | IVORY, a computer-based tool that uses cli … IVORY, a computer-based tool that uses clinical findings as the basic unit for composing progress notes, generates progress notes more efficiently than a character-based word processor. IVORY's clinical findings are contained within a structured vocabulary we developed to support generation of both prose progress notes and SNOMED III codes. Observational studies of physician participation in the development of IVORY's structured vocabulary have helped to identify areas where changes are required before IVORY will be acceptable for routine clinical use. ll be acceptable for routine clinical use. |
| A Continuous-Speech Interface to a Decision-Support System: I. Techniques to Accommodate for Misrecognized Input + | Our objective was to build a speech interf … Our objective was to build a speech interface for use with a medical decision-support application. We used a commercially available speech-recognition system to obtain textual representations of input utterances and we generated grammars to supply the anticipated utterances for the recognition system. We designed the interface to identify controlled vocabulary terms from grammatical and ungrammatical textual transcriptions of utterances. Our method included two steps: translation of the controlled vocabulary terms and textual representations of input utterances into keyword-based canonical forms, and matching canonical forms of input utterances to canonical forms of the controlled vocabulary terms. Evaluation of the speech interface indicated that this matching approach increased the identification of terms from misrecognized utterances. on of terms from misrecognized utterances. |
| A Continuous-Speech Interface to a Decision-Support System: II. An Evaluation Using a Wizard-of-Oz Experiment + | Objective: Evaluate the performance of a … Objective: Evaluate the performance of a continuous-speech interface to a decision-support system.Design: We performed a prospective evaluation of a speech interface that matches unconstrained utterances of physicians with controlled-vocabulary terms from Quick Medical Reference (QMR). We assessed the performance of the speech interface in two stages: in the real-time experiment, physician-subjects viewed audio-visual stimuli intended to evoke clinical findings, spoke a description of each finding into the speech interface, and then chose from a list generated by the interface the QMR term that most closely matched the finding. Subjects believed that the speech recognizer decoded their utterances; in reality, a hidden experimenter typed utterances into the interface (Wizard-of-Oz experimental design). Later, we replayed the same utterances through the speech recognizer and measured how accurately utterances matched with appropriate QMR terms using the results of the real-time experiment as the gold standard.Measurements: We measured how accurately the speech-recognition system converted input utterances to text strings (recognition accuracy) and how accurately the speech interface matched input utterances to appropriate QMR terms (semantic accuracy). Results: Overall recognition accuracy was less than 50%. However, using language-processing techniques that match keywords in recognized utterances to keywords in QMR terms, the semantic accuracy of the system was 81%. Conclusions: We found that reasonable semantic accuracy can be attained when language-processing techniques are used to accommodate for speech misrecognition. We also found that the Wizard-of-Oz experimental design offered many advantages for this evaluation and believe that this technique may be useful to future evaluators of speech-input systems. future evaluators of speech-input systems. |
| A Control Architecture for Intelligent Mobile Robots + | Intelligent mobile robots interacting with … Intelligent mobile robots interacting with an uncertain and dynamic environment should be able to identify their own goals and perform tasks based on goals and their interactions with the environment. In addition, they should have planning, control, and reaction capabilities spanning several orders of magnitude in reaction time and solution complexity. We present a conceptual architecture for intelligent mobile robots that adheres to these specifications. "B-Robot" is a heterogeneous architecture inspired by the gross anatomy and physiology of the mammalian central nervous system. It combines goal-and task-level reasoning and planning skills with intermediate-level servo control loops and low-level reactivity. vo control loops and low-level reactivity. |
