Daniel Berleant

MINING MEDLINE: ABSTRACTS, SENTENCES, OR PHRASES?

A growing body of works address automated mining of biochemical knowledge from digital repositories of scientific literature, such as MEDLINE. Some of these works use abstracts as the unit of text from which to extract facts. Others use sentences for this purpose, while still others use phrases. Here we compare abstracts, sentences, and phrases in MEDLINE using the standard information retrieval performance measures of recall, precision, and effectiveness, for the task of mining interactions among biochemical terms based on term co-occurrence. Results show statistically significant differences that can impact the choice of text unit.

Qualitative and quantitative simulation: bridging the gap

Abstract Shortcomings of qualitative simulation and of quantitative simulation motivate combining them to do simulations exhibiting strengths of both. The resulting class of techniques is called semiquantitative simulation . One approach to semi-quantitative simulation is to use numeric intervals to represent incomplete quantitative information. In this research we demonstrate semi-quantitative simulation using intervals in an implemented semi-quantitative simulator called Q3. Q3 progressively refines a qualitative simulation, providing increasingly specific quantitative predictions which can converge to a numerical simulation in the limit while retaining important correctness guarantees from qualitative and interval simulation techniques. Q3s simulations are based on a technique we call step size refinement . While a pure qualitative simulation has a very coarse step size, representing the state of a system trajectory at relatively few qualitatively distinct states, Q3 interpolates newly explicit states between distinct qualitative states, thereby representing more states which instantiate new constraints, leading to improved quantitative inferences. Q3s techniques have been used for prediction, measurement interpretation, diagnosis, and even analysis of the probabilities of qualitative behaviors. Because Q3 shares important expressive and inferential properties of both qualitative and quantitative simulation, Q3 helps to bridge the gap between qualitative and quantitative simulation.

Bounding the Results of Arithmetic Operations on Random Variables of Unknown Dependency Using Intervals

Many real problems involve calculations on random variables, yet precise details about the correlations or other dependency relationships among those variables are often unknown.For example consider analyzing the cancer risk associated with an environmental contaminant. The dependency of an individuals cumulative exposure on the less useful (but more obtainable) current exposure level will be uncertain. In this and many other cases, data points from which to derive such dependencies are sparse, and obtaining additional data is prohibitively expensive or difficult. Thus manipulating variables whose dependencies are unspecified is a problem of significance.This paper describes a new approach to bounding the results of arithmetic operations on random variables when the dependency relationship between the variables is unspecified. The bounds enclose the space in which the results distribution function can be.

Dependence in probabilistic modeling, Dempster-Shafer theory, and probability bounds analysis.

This report summarizes methods to incorporate information (or lack of information) about inter-variable dependence into risk assessments that use Dempster-Shafer theory or probability bounds analysis to address epistemic and aleatory uncertainty. The report reviews techniques for simulating correlated variates for a given correlation measure and dependence model, computation of bounds on distribution functions under a specified dependence model, formulation of parametric and empirical dependence models, and bounding approaches that can be used when information about the intervariable dependence is incomplete. The report also reviews several of the most pervasive and dangerous myths among risk analysts about dependence in probabilistic models.

Equivalence of methods for uncertainty propagation of real-valued random variables

Abstract In this paper we compare four methods for the reliable propagation of uncertainty through calculations involving the binary operations of addition, multiplication, subtraction and division. The methods we investigate are: (i) dependency bounds convolution; (ii) Distribution Envelope Determination; (iii) interval probabilities; and (iv) Dempster–Shafer belief functions. We show that although each of these methods were constructed for different types of applications, they converge to equivalent methods when they are restricted to cumulative distribution functions on the positive reals. We also show that while some of the methods have been formally constructed to deal only with operations on random variables under an assumption of independence, all of the methods can be extended to deal with unknown dependencies and perfect positive and negative dependence among variables.

Statool: A Tool for Distribution Envelope Determination (DEnv), an Interval-Based Algorithm for Arithmetic on Random Variables

We present Statool, a software tool for obtaining bounds on the distributions of sums, products, and various other functions of random variables where the dependency relationship of the random variables need not be specified. Statool implements the DEnv algorithm, which we have described previously [4] but not implemented. Our earlier tool addressed only the much more elementary case of independent random variables [3]. An existing tool, RiskCalc [13], also addresses the case of unknown dependency using a different algorithm [33] based on copulas [23], while descriptions and implementations of still other algorithms for similar problems will be reported soon [17] as the area proceeds through a phase of rapid development.

Version augmented URIs for reference permanencevia an Apache module design

The World Wide Web is perhaps the most convenient library resource ever to exist. Yet it also suffers from a critical flaw: WWW document citations existing as clickable links and as standard references in printed documents are quite unreliable. This arises from the ease with which these documents may be altered. Alteration of cited WWW documents may lead to the citations themselves becoming invalid or inaccurate. Printed documents are less prone to such problems due to the greater difficulty in altering or eradicating them. As a consequence of the alterability of WWW documents, the whole traditional system of new work building upon an archive of unchangeable previous work ceases to be valid. For the first time in history, documents are being built upon a shifting foundation to an extent of ever increasing and serious proportions. This work investigates a system aimed at encouraging the stability of WWW documents. In doing so it helps to preserve the accuracy of citations over time. The system presented extends a documents URI (Universal Resource Identifier) to include date or revision information thus allowing document providers to permit users to refer accurately to a particular document version. This paper both describes the system and demonstrates its use. The system is implemented as an Apache WWW server module. The module processes and retrieves documents referenced with extended URIs. It provides a model public domain system for alleviating the problem of content stable WWW documents.

Extracting biochemical interactions from MEDLINE using a link grammar parser

Many natural language processing approaches at various complexity levels have been reported for extracting biochemical interactions from MEDLINE. While some algorithms using simple template matching are unable to deal with the complex syntactic structures, others exploiting sophisticated parsing techniques are hindered by greater computational cost. This study investigates link grammar parsing for extracting biochemical interactions. Link grammar parsing can handle many syntactic structures and is computationally relatively efficient. We experimented on a sample MEDLINE corpus. Although the parser was originally developed for conversational English and made many mistakes in parsing sentences from the biochemical domain, it nevertheless achieved better overall performance than a co-occurrence-only method. Customizing the parser for the biomedical domain is expected to improve its performance further.

Representation and problem solving with Distribution Envelope Determination (DEnv)

Distribution Envelope Determination (DEnv) is a method for computing the CDFs of random variables whose samples are a function of samples of other random variable(s), termed inputs. DEnv computes envelopes around these CDFs when there is uncertainty about the precise form of the probability distribution describing any input. For example, inputs whose distribution functions have means and variances known only to within intervals can be handled. More generally, inputs can be handled if the set of all plausible cumulative distributions describing each input can be enclosed between left and right envelopes. Results will typically be in the form of envelopes when inputs are envelopes, when the dependency relationship of the inputs is unspecified, or both. For example in the case of specific input distribution functions with unspecified dependency relationships, each of the infinite number of possible dependency relationships would imply some specific output distribution, and the set of all such output distributions can be bounded with envelopes. The DEnv algorithm is a way to obtain the bounding envelopes. DEnv is implemented in a tool which is used to solve problems from a benchmark set.

Automatically Verified Arithmetic on Probability Distributions and Intervals

In this chapter we address two related problems: 1. Representing and operating on operands which are probability distribution functions; and 2. Representing and operating on operands when one is a distribution function and the other is an interval.