id
stringlengths
9
16
title
stringlengths
4
278
categories
stringlengths
5
104
abstract
stringlengths
6
4.09k
1302.6667
Crowdsourcing for Bioinformatics
q-bio.QM cs.CY cs.SI physics.soc-ph
Motivation: Bioinformatics is faced with a variety of problems that require human involvement. Tasks like genome annotation, image analysis, knowledge-base construction and protein structure determination all benefit from human input. In some cases people are needed in vast quantities while in others we need just a few with very rare abilities. Crowdsourcing encompasses an emerging collection of approaches for harnessing such distributed human intelligence. Recently, the bioinformatics community has begun to apply crowdsourcing in a variety of contexts, yet few resources are available that describe how these human-powered systems work and how to use them effectively in scientific domains. Results: Here, we provide a framework for understanding and applying several different types of crowdsourcing. The framework considers two broad classes: systems for solving large-volume 'microtasks' and systems for solving high-difficulty 'megatasks'. Within these classes, we discuss system types including: volunteer labor, games with a purpose, microtask markets and open innovation contests. We illustrate each system type with successful examples in bioinformatics and conclude with a guide for matching problems to crowdsourcing solutions.
1302.6668
Finite-time consensus using stochastic matrices with positive diagonals
cs.MA cs.SY
We discuss the possibility of reaching consensus in finite time using only linear iterations, with the additional restrictions that the update matrices must be stochastic with positive diagonals and consistent with a given graph structure. We show that finite-time average consensus can always be achieved for connected undirected graphs. For directed graphs, we show some necessary conditions for finite-time consensus, including strong connectivity and the presence of a simple cycle of even length.
1302.6677
Taming the Curse of Dimensionality: Discrete Integration by Hashing and Optimization
cs.LG cs.AI stat.ML
Integration is affected by the curse of dimensionality and quickly becomes intractable as the dimensionality of the problem grows. We propose a randomized algorithm that, with high probability, gives a constant-factor approximation of a general discrete integral defined over an exponentially large set. This algorithm relies on solving only a small number of instances of a discrete combinatorial optimization problem subject to randomly generated parity constraints used as a hash function. As an application, we demonstrate that with a small number of MAP queries we can efficiently approximate the partition function of discrete graphical models, which can in turn be used, for instance, for marginal computation or model selection.
1302.6683
Decentralized set-valued state estimation based on non-deterministic chains
cs.SY
A general decentralized computational framework for set-valued state estimation and prediction for the class of systems that accept a hybrid state machine representation is considered in this article. The decentralized scheme consists of a conjunction of distributed state machines that are specified by a decomposition of the external signal space. While this is shown to produce, in general, outer approximations of the outcomes of the original monolithic state machine, here, specific rules for the signal space decomposition are devised by utilizing structural properties of the underyling transition relation, leading to a recovery of the exact state set results. By applying a suitable approximation algorithm, we show that computational complexity in the decentralized setting may thereby essentially reduce as compared to the centralized estimation scheme.
1302.6703
Compressive Sensing for Spread Spectrum Receivers
cs.IT math.IT
With the advent of ubiquitous computing there are two design parameters of wireless communication devices that become very important power: efficiency and production cost. Compressive sensing enables the receiver in such devices to sample below the Shannon-Nyquist sampling rate, which may lead to a decrease in the two design parameters. This paper investigates the use of Compressive Sensing (CS) in a general Code Division Multiple Access (CDMA) receiver. We show that when using spread spectrum codes in the signal domain, the CS measurement matrix may be simplified. This measurement scheme, named Compressive Spread Spectrum (CSS), allows for a simple, effective receiver design. Furthermore, we numerically evaluate the proposed receiver in terms of bit error rate under different signal to noise ratio conditions and compare it with other receiver structures. These numerical experiments show that though the bit error rate performance is degraded by the subsampling in the CS-enabled receivers, this may be remedied by including quantization in the receiver model. We also study the computational complexity of the proposed receiver design under different sparsity and measurement ratios. Our work shows that it is possible to subsample a CDMA signal using CSS and that in one example the CSS receiver outperforms the classical receiver.
1302.6704
Decentralized set-valued state estimation and prediction for hybrid systems: A symbolic approach
cs.SY
A symbolic approach to decentralized set-valued state estimation and prediction for systems that admit a hybrid state machine representations is proposed. The decentralized computational scheme represents a conj unction of a finite number of distributed state machines, which are specified by an appropriate decomposition of the external signal space. It aims at a distribution of computational tasks into smaller ones, allocated to individual distributed state machines, leading to a potentially significant reduction in the overall space/time computational complexity. We show that, in general, such a scheme outerapproximates the state set estimates and predictions of the original monolithic state machine. By utilizing structural properties of the transition relation of the latter, in a next step, we propose constructive decomposition algorithms for a recovery of the exact state set outcomes.
1302.6738
Finding overlapping communities in networks using evolutionary method
cs.SI physics.soc-ph
Community structure is a typical property of many real-world networks, and has become a key to understand the dynamics of the networked systems. In these networks most nodes apparently lie in a community while there often exists a few nodes straddling several communities. An ideal algorithm for community detection is preferable which can identify the overlapping communities in such networks. To represent an overlapping division we develop a encoding schema composed of two segments, the first one represents a disjoint partition and the second one represents a extension of the partition that allows of multiple memberships. We give a measure for the informativeness of a node, and present an evolutionary method for detecting the overlapping communities in a network.
1302.6764
Categorizing Bugs with Social Networks: A Case Study on Four Open Source Software Communities
cs.SE cs.LG cs.SI nlin.AO physics.soc-ph
Efficient bug triaging procedures are an important precondition for successful collaborative software engineering projects. Triaging bugs can become a laborious task particularly in open source software (OSS) projects with a large base of comparably inexperienced part-time contributors. In this paper, we propose an efficient and practical method to identify valid bug reports which a) refer to an actual software bug, b) are not duplicates and c) contain enough information to be processed right away. Our classification is based on nine measures to quantify the social embeddedness of bug reporters in the collaboration network. We demonstrate its applicability in a case study, using a comprehensive data set of more than 700,000 bug reports obtained from the Bugzilla installation of four major OSS communities, for a period of more than ten years. For those projects that exhibit the lowest fraction of valid bug reports, we find that the bug reporters' position in the collaboration network is a strong indicator for the quality of bug reports. Based on this finding, we develop an automated classification scheme that can easily be integrated into bug tracking platforms and analyze its performance in the considered OSS communities. A support vector machine (SVM) to identify valid bug reports based on the nine measures yields a precision of up to 90.3% with an associated recall of 38.9%. With this, we significantly improve the results obtained in previous case studies for an automated early identification of bugs that are eventually fixed. Furthermore, our study highlights the potential of using quantitative measures of social organization in collaborative software engineering. It also opens a broad perspective for the integration of social awareness in the design of support infrastructures.
1302.6768
Missing Entries Matrix Approximation and Completion
math.NA cs.LG stat.ML
We describe several algorithms for matrix completion and matrix approximation when only some of its entries are known. The approximation constraint can be any whose approximated solution is known for the full matrix. For low rank approximations, similar algorithms appears recently in the literature under different names. In this work, we introduce new theorems for matrix approximation and show that these algorithms can be extended to handle different constraints such as nuclear norm, spectral norm, orthogonality constraints and more that are different than low rank approximations. As the algorithms can be viewed from an optimization point of view, we discuss their convergence to global solution for the convex case. We also discuss the optimal step size and show that it is fixed in each iteration. In addition, the derived matrix completion flow is robust and does not require any parameters. This matrix completion flow is applicable to different spectral minimizations and can be applied to physics, mathematics and electrical engineering problems such as data reconstruction of images and data coming from PDEs such as Helmholtz equation used for electromagnetic waves.
1302.6770
Total communicability as a centrality measure
cs.SI math.NA physics.soc-ph
We examine a node centrality measure based on the notion of total communicability, defined in terms of the row sums of the exponential of the adjacency matrix of the network. We argue that this is a natural metric for ranking nodes in a network, and we point out that it can be computed very rapidly even in the case of large networks. Furthermore, we propose the total sum of node communicabilities as a useful measure of network connectivity. Extensive numerical studies are conducted in order to compare this centrality measure with the closely related ones of subgraph centrality [E. Estrada and J. A. Rodriguez-Velazquez, Phys. Rev. E, 71 (2005), 056103] and Katz centrality [L. Katz, Psychometrica, 18 (1953), pp. 39-43]. Both synthetic and real-world networks are used in the computations.
1302.6777
Ending-based Strategies for Part-of-speech Tagging
cs.CL
Probabilistic approaches to part-of-speech tagging rely primarily on whole-word statistics about word/tag combinations as well as contextual information. But experience shows about 4 per cent of tokens encountered in test sets are unknown even when the training set is as large as a million words. Unseen words are tagged using secondary strategies that exploit word features such as endings, capitalizations and punctuation marks. In this work, word-ending statistics are primary and whole-word statistics are secondary. First, a tagger was trained and tested on word endings only. Subsequent experiments added back whole-word statistics for the words occurring most frequently in the training set. As grew larger, performance was expected to improve, in the limit performing the same as word-based taggers. Surprisingly, the ending-based tagger initially performed nearly as well as the word-based tagger; in the best case, its performance significantly exceeded that of the word-based tagger. Lastly, and unexpectedly, an effect of negative returns was observed - as grew larger, performance generally improved and then declined. By varying factors such as ending length and tag-list strategy, we achieved a success rate of 97.5 percent.
1302.6779
An Evaluation of an Algorithm for Inductive Learning of Bayesian Belief Networks Usin
cs.AI
Bayesian learning of belief networks (BLN) is a method for automatically constructing belief networks (BNs) from data using search and Bayesian scoring techniques. K2 is a particular instantiation of the method that implements a greedy search strategy. To evaluate the accuracy of K2, we randomly generated a number of BNs and for each of those we simulated data sets. K2 was then used to induce the generating BNs from the simulated data. We examine the performance of the program, and the factors that influence it. We also present a simple BN model, developed from our results, which predicts the accuracy of K2, when given various characteristics of the data set.
1302.6780
Probabilistic Constraint Satisfaction with Non-Gaussian Noise
cs.AI
We have previously reported a Bayesian algorithm for determining the coordinates of points in three-dimensional space from uncertain constraints. This method is useful in the determination of biological molecular structure. It is limited, however, by the requirement that the uncertainty in the constraints be normally distributed. In this paper, we present an extension of the original algorithm that allows constraint uncertainty to be represented as a mixture of Gaussians, and thereby allows arbitrary constraint distributions. We illustrate the performance of this algorithm on a problem drawn from the domain of molecular structure determination, in which a multicomponent constraint representation produces a much more accurate solution than the old single component mechanism. The new mechanism uses mixture distributions to decompose the problem into a set of independent problems with unimodal constraint uncertainty. The results of the unimodal subproblems are periodically recombined using Bayes' law, to avoid combinatorial explosion. The new algorithm is particularly suited for parallel implementation.
1302.6781
A Bayesian Method Reexamined
cs.AI
This paper examines the "K2" network scoring metric of Cooper and Herskovits. It shows counterintuitive results from applying this metric to simple networks. One family of noninformative priors is suggested for assigning equal scores to equivalent networks.
1302.6782
Laplace's Method Approximations for Probabilistic Inference in Belief Networks with Continuous Variables
cs.AI
Laplace's method, a family of asymptotic methods used to approximate integrals, is presented as a potential candidate for the tool box of techniques used for knowledge acquisition and probabilistic inference in belief networks with continuous variables. This technique approximates posterior moments and marginal posterior distributions with reasonable accuracy [errors are O(n^-2) for posterior means] in many interesting cases. The method also seems promising for computing approximations for Bayes factors for use in the context of model selection, model uncertainty and mixtures of pdfs. The limitations, regularity conditions and computational difficulties for the implementation of Laplace's method are comparable to those associated with the methods of maximum likelihood and posterior mode analysis.
1302.6783
Generating New Beliefs From Old
cs.AI
In previous work [BGHK92, BGHK93], we have studied the random-worlds approach -- a particular (and quite powerful) method for generating degrees of belief (i.e., subjective probabilities) from a knowledge base consisting of objective (first-order, statistical, and default) information. But allowing a knowledge base to contain only objective information is sometimes limiting. We occasionally wish to include information about degrees of belief in the knowledge base as well, because there are contexts in which old beliefs represent important information that should influence new beliefs. In this paper, we describe three quite general techniques for extending a method that generates degrees of belief from objective information to one that can make use of degrees of belief as well. All of our techniques are bloused on well-known approaches, such as cross-entropy. We discuss general connections between the techniques and in particular show that, although conceptually and technically quite different, all of the techniques give the same answer when applied to the random-worlds method.
1302.6784
Counterfactual Probabilities: Computational Methods, Bounds and Applications
cs.AI
Evaluation of counterfactual queries (e.g., "If A were true, would C have been true?") is important to fault diagnosis, planning, and determination of liability. In this paper we present methods for computing the probabilities of such queries using the formulation proposed in [Balke and Pearl, 1994], where the antecedent of the query is interpreted as an external action that forces the proposition A to be true. When a prior probability is available on the causal mechanisms governing the domain, counterfactual probabilities can be evaluated precisely. However, when causal knowledge is specified as conditional probabilities on the observables, only bounds can computed. This paper develops techniques for evaluating these bounds, and demonstrates their use in two applications: (1) the determination of treatment efficacy from studies in which subjects may choose their own treatment, and (2) the determination of liability in product-safety litigation.
1302.6786
Modus Ponens Generating Function in the Class of ^-valuations of Plausibility
cs.AI
We discuss the problem of construction of inference procedures which can manipulate with uncertainties measured in ordinal scales and fulfill to the property of strict monotonicity of conclusion. The class of A-valuations of plausibility is considered where operations based only on information about linear ordering of plausibility values are used. In this class the modus ponens generating function fulfiling to the property of strict monotonicity of conclusions is introduced.
1302.6787
Approximation Algorithms for the Loop Cutset Problem
cs.AI cs.DS
We show how to find a small loop curser in a Bayesian network. Finding such a loop cutset is the first step in the method of conditioning for inference. Our algorithm for finding a loop cutset, called MGA, finds a loop cutset which is guaranteed in the worst case to contain less than twice the number of variables contained in a minimum loop cutset. We test MGA on randomly generated graphs and find that the average ratio between the number of instances associated with the algorithms' output and the number of instances associated with a minimum solution is 1.22.
1302.6788
Possibility and Necessity Functions over Non-classical Logics
cs.AI
We propose an integration of possibility theory into non-classical logics. We obtain many formal results that generalize the case where possibility and necessity functions are based on classical logic. We show how useful such an approach is by applying it to reasoning under uncertain and inconsistent information.
1302.6789
Exploratory Model Building
cs.AI
Some instances of creative thinking require an agent to build and test hypothetical theories. Such a reasoner needs to explore the space of not only those situations that have occurred in the past, but also those that are rationally conceivable. In this paper we present a formalism for exploring the space of conceivable situation-models for those domains in which the knowledge is primarily probabilistic in nature. The formalism seeks to construct consistent, minimal, and desirable situation-descriptions by selecting suitable domain-attributes and dependency relationships from the available domain knowledge.
1302.6791
Planning with External Events
cs.AI
I describe a planning methodology for domains with uncertainty in the form of external events that are not completely predictable. The events are represented by enabling conditions and probabilities of occurrence. The planner is goal-directed and backward chaining, but the subgoals are suggested by analyzing the probability of success of the partial plan rather than being simply the open conditions of the operators in the plan. The partial plan is represented as a Bayesian belief net to compute its probability of success. Since calculating the probability of success of a plan can be very expensive I introduce two other techniques for computing it, one that uses Monte Carlo simulation to estimate it and one based on a Markov chain representation that uses knowledge about the dependencies between the predicates describing the domain.
1302.6792
Properties of Bayesian Belief Network Learning Algorithms
cs.AI
Bayesian belief network learning algorithms have three basic components: a measure of a network structure and a database, a search heuristic that chooses network structures to be considered, and a method of estimating the probability tables from the database. This paper contributes to all these three topics. The behavior of the Bayesian measure of Cooper and Herskovits and a minimum description length (MDL) measure are compared with respect to their properties for both limiting size and finite size databases. It is shown that the MDL measure has more desirable properties than the Bayesian measure when a distribution is to be learned. It is shown that selecting belief networks with certain minimallity properties is NP-hard. This result justifies the use of search heuristics instead of exact algorithms for choosing network structures to be considered. In some cases, a collection of belief networks can be represented by a single belief network which leads to a new kind of probability table estimation called smoothing. We argue that smoothing can be efficiently implemented by incorporating it in the search heuristic. Experimental results suggest that for learning probabilities of belief networks smoothing is helpful.
1302.6793
A Stratified Simulation Scheme for Inference in Bayesian Belief Networks
cs.AI
Simulation schemes for probabilistic inference in Bayesian belief networks offer many advantages over exact algorithms; for example, these schemes have a linear and thus predictable runtime while exact algorithms have exponential runtime. Experiments have shown that likelihood weighting is one of the most promising simulation schemes. In this paper, we present a new simulation scheme that generates samples more evenly spread in the sample space than the likelihood weighting scheme. We show both theoretically and experimentally that the stratified scheme outperforms likelihood weighting in average runtime and error in estimates of beliefs.
1302.6794
Efficient Estimation of the Value of Information in Monte Carlo Models
cs.AI
The expected value of information (EVI) is the most powerful measure of sensitivity to uncertainty in a decision model: it measures the potential of information to improve the decision, and hence measures the expected value of outcome. Standard methods for computing EVI use discrete variables and are computationally intractable for models that contain more than a few variables. Monte Carlo simulation provides the basis for more tractable evaluation of large predictive models with continuous and discrete variables, but so far computation of EVI in a Monte Carlo setting also has appeared impractical. We introduce an approximate approach based on pre-posterior analysis for estimating EVI in Monte Carlo models. Our method uses a linear approximation to the value function and multiple linear regression to estimate the linear model from the samples. The approach is efficient and practical for extremely large models. It allows easy estimation of EVI for perfect or partial information on individual variables or on combinations of variables. We illustrate its implementation within Demos (a decision modeling system), and its application to a large model for crisis transportation planning.
1302.6795
Symbolic Probabilitistic Inference in Large BN2O Networks
cs.AI
A BN2O network is a two level belief net in which the parent interactions are modeled using the noisy-or interaction model. In this paper we discuss application of the SPI local expression language to efficient inference in large BN2O networks. In particular, we show that there is significant structure, which can be exploited to improve over the Quickscore result. We further describe how symbolic techniques can provide information which can significantly reduce the computation required for computing all cause posterior marginals. Finally, we present a novel approximation technique with preliminary experimental results.
1302.6796
Action Networks: A Framework for Reasoning about Actions and Change under Uncertainty
cs.AI
This work proposes action networks as a semantically well-founded framework for reasoning about actions and change under uncertainty. Action networks add two primitives to probabilistic causal networks: controllable variables and persistent variables. Controllable variables allow the representation of actions as directly setting the value of specific events in the domain, subject to preconditions. Persistent variables provide a canonical model of persistence according to which both the state of a variable and the causal mechanism dictating its value persist over time unless intervened upon by an action (or its consequences). Action networks also allow different methods for quantifying the uncertainty in causal relationships, which go beyond traditional probabilistic quantification. This paper describes both recent results and work in progress.
1302.6797
On the Relation between Kappa Calculus and Probabilistic Reasoning
cs.AI
We study the connection between kappa calculus and probabilistic reasoning in diagnosis applications. Specifically, we abstract a probabilistic belief network for diagnosing faults into a kappa network and compare the ordering of faults computed using both methods. We show that, at least for the example examined, the ordering of faults coincide as long as all the causal relations in the original probabilistic network are taken into account. We also provide a formal analysis of some network structures where the two methods will differ. Both kappa rankings and infinitesimal probabilities have been used extensively to study default reasoning and belief revision. But little has been done on utilizing their connection as outlined above. This is partly because the relation between kappa and probability calculi assumes that probabilities are arbitrarily close to one (or zero). The experiments in this paper investigate this relation when this assumption is not satisfied. The reported results have important implications on the use of kappa rankings to enhance the knowledge engineering of uncertainty models.
1302.6798
A Structured, Probabilistic Representation of Action
cs.AI
When agents devise plans for execution in the real world, they face two important forms of uncertainty: they can never have complete knowledge about the state of the world, and they do not have complete control, as the effects of their actions are uncertain. While most classical planning methods avoid explicit uncertainty reasoning, we believe that uncertainty should be explicitly represented and reasoned about. We develop a probabilistic representation for states and actions, based on belief networks. We define conditional belief nets (CBNs) to capture the probabilistic dependency of the effects of an action upon the state of the world. We also use a CBN to represent the intrinsic relationships among entities in the environment, which persist from state to state. We present a simple projection algorithm to construct the belief network of the state succeeding an action, using the environment CBN model to infer indirect effects. We discuss how the qualitative aspects of belief networks and CBNs make them appropriate for the various stages of the problem solving process, from model construction to the design of planning algorithms.
1302.6799
Integrating Planning and Execution in Stochastic Domains
cs.AI
We investigate planning in time-critical domains represented as Markov Decision Processes, showing that search based techniques can be a very powerful method for finding close to optimal plans. To reduce the computational cost of planning in these domains, we execute actions as we construct the plan, and sacrifice optimality by searching to a fixed depth and using a heuristic function to estimate the value of states. Although this paper concentrates on the search algorithm, we also discuss ways of constructing heuristic functions suitable for this approach. Our results show that by interleaving search and execution, close to optimal policies can be found without the computational requirements of other approaches.
1302.6800
Localized Partial Evaluation of Belief Networks
cs.AI
Most algorithms for propagating evidence through belief networks have been exact and exhaustive: they produce an exact (point-valued) marginal probability for every node in the network. Often, however, an application will not need information about every n ode in the network nor will it need exact probabilities. We present the localized partial evaluation (LPE) propagation algorithm, which computes interval bounds on the marginal probability of a specified query node by examining a subset of the nodes in the entire network. Conceptually, LPE ignores parts of the network that are "too far away" from the queried node to have much impact on its value. LPE has the "anytime" property of being able to produce better solutions (tighter intervals) given more time to consider more of the network.
1302.6801
A Probabilistic Model of Action for Least-Commitment Planning with Information Gather
cs.AI
AI planning algorithms have addressed the problem of generating sequences of operators that achieve some input goal, usually assuming that the planning agent has perfect control over and information about the world. Relaxing these assumptions requires an extension to the action representation that allows reasoning both about the changes an action makes and the information it provides. This paper presents an action representation that extends the deterministic STRIPS model, allowing actions to have both causal and informational effects, both of which can be context dependent and noisy. We also demonstrate how a standard least-commitment planning algorithm can be extended to include informational actions and contingent execution.
1302.6802
Some Properties of Joint Probability Distributions
cs.AI
Several Artificial Intelligence schemes for reasoning under uncertainty explore either explicitly or implicitly asymmetries among probabilities of various states of their uncertain domain models. Even though the correct working of these schemes is practically contingent upon the existence of a small number of probable states, no formal justification has been proposed of why this should be the case. This paper attempts to fill this apparent gap by studying asymmetries among probabilities of various states of uncertain models. By rewriting the joint probability distribution over a model's variables into a product of individual variables' prior and conditional probability distributions, and applying central limit theorem to this product, we can demonstrate that the probabilities of individual states of the model can be expected to be drawn from highly skewed, log-normal distributions. With sufficient asymmetry in individual prior and conditional probability distributions, a small fraction of states can be expected to cover a large portion of the total probability space with the remaining states having practically negligible probability. Theoretical discussion is supplemented by simulation results and an illustrative real-world example.
1302.6803
An Ordinal View of Independence with Application to Plausible Reasoning
cs.AI
An ordinal view of independence is studied in the framework of possibility theory. We investigate three possible definitions of dependence, of increasing strength. One of them is the counterpart to the multiplication law in probability theory, and the two others are based on the notion of conditional possibility. These two have enough expressive power to support the whole possibility theory, and a complete axiomatization is provided for the strongest one. Moreover we show that weak independence is well-suited to the problems of belief change and plausible reasoning, especially to address the problem of blocking of property inheritance in exception-tolerant taxonomic reasoning.
1302.6804
Penalty logic and its Link with Dempster-Shafer Theory
cs.AI
Penalty logic, introduced by Pinkas, associates to each formula of a knowledge base the price to pay if this formula is violated. Penalties may be used as a criterion for selecting preferred consistent subsets in an inconsistent knowledge base, thus inducing a non-monotonic inference relation. A precise formalization and the main properties of penalty logic and of its associated non-monotonic inference relation are given in the first part. We also show that penalty logic and Dempster-Shafer theory are related, especially in the infinitesimal case.
1302.6805
Value of Evidence on Influence Diagrams
cs.AI
In this paper, we introduce evidence propagation operations on influence diagrams and a concept of value of evidence, which measures the value of experimentation. Evidence propagation operations are critical for the computation of the value of evidence, general update and inference operations in normative expert systems which are based on the influence diagram (generalized Bayesian network) paradigm. The value of evidence allows us to compute directly an outcome sensitivity, a value of perfect information and a value of control which are used in decision analysis (the science of decision making under uncertainty). More specifically, the outcome sensitivity is the maximum difference among the values of evidence, the value of perfect information is the expected value of the values of evidence, and the value of control is the optimal value of the values of evidence. We also discuss an implementation and a relative computational efficiency issues related to the value of evidence and the value of perfect information.
1302.6806
Conditional Independence in Possibility Theory
cs.AI
Possibilistic conditional independence is investigated: we propose a definition of this notion similar to the one used in probability theory. The links between independence and non-interactivity are investigated, and properties of these relations are given. The influence of the conjunction used to define a conditional measure of possibility is also highlighted: we examine three types of conjunctions: Lukasiewicz - like T-norms, product-like T-norms and the minimum operator.
1302.6807
Backward Simulation in Bayesian Networks
cs.AI
Backward simulation is an approximate inference technique for Bayesian belief networks. It differs from existing simulation methods in that it starts simulation from the known evidence and works backward (i.e., contrary to the direction of the arcs). The technique's focus on the evidence leads to improved convergence in situations where the posterior beliefs are dominated by the evidence rather than by the prior probabilities. Since this class of situations is large, the technique may make practical the application of approximate inference in Bayesian belief networks to many real-world problems.
1302.6808
Learning Gaussian Networks
cs.AI cs.LG stat.ML
We describe algorithms for learning Bayesian networks from a combination of user knowledge and statistical data. The algorithms have two components: a scoring metric and a search procedure. The scoring metric takes a network structure, statistical data, and a user's prior knowledge, and returns a score proportional to the posterior probability of the network structure given the data. The search procedure generates networks for evaluation by the scoring metric. Previous work has concentrated on metrics for domains containing only discrete variables, under the assumption that data represents a multinomial sample. In this paper, we extend this work, developing scoring metrics for domains containing all continuous variables or a mixture of discrete and continuous variables, under the assumption that continuous data is sampled from a multivariate normal distribution. Our work extends traditional statistical approaches for identifying vanishing regression coefficients in that we identify two important assumptions, called event equivalence and parameter modularity, that when combined allow the construction of prior distributions for multivariate normal parameters from a single prior Bayesian network specified by a user.
1302.6809
On Testing Whether an Embedded Bayesian Network Represents a Probability Model
cs.AI
Testing the validity of probabilistic models containing unmeasured (hidden) variables is shown to be a hard task. We show that the task of testing whether models are structurally incompatible with the data at hand, requires an exponential number of independence evaluations, each of the form: "X is conditionally independent of Y, given Z." In contrast, a linear number of such evaluations is required to test a standard Bayesian network (one per vertex). On the positive side, we show that if a network with hidden variables G has a tree skeleton, checking whether G represents a given probability model P requires the polynomial number of such independence evaluations. Moreover, we provide an algorithm that efficiently constructs a tree-structured Bayesian network (with hidden variables) that represents P if such a network exists, and further recognizes when such a network does not exist.
1302.6810
Epsilon-Safe Planning
cs.AI
We introduce an approach to high-level conditional planning we call epsilon-safe planning. This probabilistic approach commits us to planning to meet some specified goal with a probability of success of at least 1-epsilon for some user-supplied epsilon. We describe several algorithms for epsilon-safe planning based on conditional planners. The two conditional planners we discuss are Peot and Smith's nonlinear conditional planner, CNLP, and our own linear conditional planner, PLINTH. We present a straightforward extension to conditional planners for which computing the necessary probabilities is simple, employing a commonly-made but perhaps overly-strong independence assumption. We also discuss a second approach to epsilon-safe planning which relaxes this independence assumption, involving the incremental construction of a probability dependence model in conjunction with the construction of the plan graph.
1302.6811
Generating Bayesian Networks from Probability Logic Knowledge Bases
cs.AI
We present a method for dynamically generating Bayesian networks from knowledge bases consisting of first-order probability logic sentences. We present a subset of probability logic sufficient for representing the class of Bayesian networks with discrete-valued nodes. We impose constraints on the form of the sentences that guarantee that the knowledge base contains all the probabilistic information necessary to generate a network. We define the concept of d-separation for knowledge bases and prove that a knowledge base with independence conditions defined by d-separation is a complete specification of a probability distribution. We present a network generation algorithm that, given an inference problem in the form of a query Q and a set of evidence E, generates a network to compute P(Q|E). We prove the algorithm to be correct.
1302.6812
Abstracting Probabilistic Actions
cs.AI
This paper discusses the problem of abstracting conditional probabilistic actions. We identify two distinct types of abstraction: intra-action abstraction and inter-action abstraction. We define what it means for the abstraction of an action to be correct and then derive two methods of intra-action abstraction and two methods of inter-action abstraction which are correct according to this criterion. We illustrate the developed techniques by applying them to actions described with the temporal action representation used in the DRIPS decision-theoretic planner and we describe how the planner uses abstraction to reduce the complexity of planning.
1302.6813
On Modal Logics for Qualitative Possibility in a Fuzzy Setting
cs.LO cs.AI
Within the possibilistic approach to uncertainty modeling, the paper presents a modal logical system to reason about qualitative (comparative) statements of the possibility (and necessity) of fuzzy propositions. We relate this qualitative modal logic to the many--valued analogues MVS5 and MVKD45 of the well known modal logics of knowledge and belief S5 and KD45 respectively. Completeness results are obtained for such logics and therefore, they extend previous existing results for qualitative possibilistic logics in the classical non-fuzzy setting.
1302.6814
A New Look at Causal Independence
cs.AI
Heckerman (1993) defined causal independence in terms of a set of temporal conditional independence statements. These statements formalized certain types of causal interaction where (1) the effect is independent of the order that causes are introduced and (2) the impact of a single cause on the effect does not depend on what other causes have previously been applied. In this paper, we introduce an equivalent a temporal characterization of causal independence based on a functional representation of the relationship between causes and the effect. In this representation, the interaction between causes and effect can be written as a nested decomposition of functions. Causal independence can be exploited by representing this decomposition in the belief network, resulting in representations that are more efficient for inference than general causal models. We present empirical results showing the benefits of a causal-independence representation for belief-network inference.
1302.6815
Learning Bayesian Networks: The Combination of Knowledge and Statistical Data
cs.AI
We describe algorithms for learning Bayesian networks from a combination of user knowledge and statistical data. The algorithms have two components: a scoring metric and a search procedure. The scoring metric takes a network structure, statistical data, and a user's prior knowledge, and returns a score proportional to the posterior probability of the network structure given the data. The search procedure generates networks for evaluation by the scoring metric. Our contributions are threefold. First, we identify two important properties of metrics, which we call event equivalence and parameter modularity. These properties have been mostly ignored, but when combined, greatly simplify the encoding of a user's prior knowledge. In particular, a user can express her knowledge-for the most part-as a single prior Bayesian network for the domain. Second, we describe local search and annealing algorithms to be used in conjunction with scoring metrics. In the special case where each node has at most one parent, we show that heuristic search can be replaced with a polynomial algorithm to identify the networks with the highest score. Third, we describe a methodology for evaluating Bayesian-network learning algorithms. We apply this approach to a comparison of metrics and search procedures.
1302.6816
A Decision-Based View of Causality
cs.AI
Most traditional models of uncertainty have focused on the associational relationship among variables as captured by conditional dependence. In order to successfully manage intelligent systems for decision making, however, we must be able to predict the effects of actions. In this paper, we attempt to unite two branches of research that address such predictions: causal modeling and decision analysis. First, we provide a definition of causal dependence in decision-analytic terms, which we derive from consequences of causal dependence cited in the literature. Using this definition, we show how causal dependence can be represented within an influence diagram. In particular, we identify two inadequacies of an ordinary influence diagram as a representation for cause. We introduce a special class of influence diagrams, called causal influence diagrams, which corrects one of these problems, and identify situations where the other inadequacy can be eliminated. In addition, we describe the relationships between Howard Canonical Form and existing graphical representations of cause.
1302.6817
Probabilistic Description Logics
cs.AI
On the one hand, classical terminological knowledge representation excludes the possibility of handling uncertain concept descriptions involving, e.g., "usually true" concept properties, generalized quantifiers, or exceptions. On the other hand, purely numerical approaches for handling uncertainty in general are unable to consider terminological knowledge. This paper presents the language ACP which is a probabilistic extension of terminological logics and aims at closing the gap between the two areas of research. We present the formal semantics underlying the language ALUP and introduce the probabilistic formalism that is based on classes of probabilities and is realized by means of probabilistic constraints. Besides inferring implicitly existent probabilistic relationships, the constraints guarantee terminological and probabilistic consistency. Altogether, the new language ALUP applies to domains where both term descriptions and uncertainty have to be handled.
1302.6818
An Experimental Comparison of Numerical and Qualitative Probabilistic Reasoning
cs.AI
Qualitative and infinitesimal probability schemes are consistent with the axioms of probability theory, but avoid the need for precise numerical probabilities. Using qualitative probabilities could substantially reduce the effort for knowledge engineering and improve the robustness of results. We examine experimentally how well infinitesimal probabilities (the kappa-calculus of Goldszmidt and Pearl) perform a diagnostic task - troubleshooting a car that will not start by comparison with a conventional numerical belief network. We found the infinitesimal scheme to be as good as the numerical scheme in identifying the true fault. The performance of the infinitesimal scheme worsens significantly for prior fault probabilities greater than 0.03. These results suggest that infinitesimal probability methods may be of substantial practical value for machine diagnosis with small prior fault probabilities.
1302.6819
An Alternative Proof Method for Possibilistic Logic and its Application to Terminological Logics
cs.AI
Possibilistic logic, an extension of first-order logic, deals with uncertainty that can be estimated in terms of possibility and necessity measures. Syntactically, this means that a first-order formula is equipped with a possibility degree or a necessity degree that expresses to what extent the formula is possibly or necessarily true. Possibilistic resolution yields a calculus for possibilistic logic which respects the semantics developed for possibilistic logic. A drawback, which possibilistic resolution inherits from classical resolution, is that it may not terminate if applied to formulas belonging to decidable fragments of first-order logic. Therefore we propose an alternative proof method for possibilistic logic. The main feature of this method is that it completely abstracts from a concrete calculus but uses as basic operation a test for classical entailment. We then instantiate possibilistic logic with a terminological logic, which is a decidable subclass o f first-order logic but nevertheless much more expressive than propositional logic. This yields an extension of terminological logics towards the representation of uncertain knowledge which is satisfactory from a semantic as well as algorithmic point of view.
1302.6820
Possibilistic Conditioning and Propagation
cs.AI
We give an axiomatization of confidence transfer - a known conditioning scheme - from the perspective of expectation-based inference in the sense of Gardenfors and Makinson. Then, we use the notion of belief independence to "filter out" different proposal s of possibilistic conditioning rules, all are variations of confidence transfer. Among the three rules that we consider, only Dempster's rule of conditioning passes the test of supporting the notion of belief independence. With the use of this conditioning rule, we then show that we can use local computation for computing desired conditional marginal possibilities of the joint possibility satisfying the given constraints. It turns out that our local computation scheme is already proposed by Shenoy. However, our intuitions are completely different from that of Shenoy. While Shenoy just defines a local computation scheme that fits his framework of valuation-based systems, we derive that local computation scheme from II(,8) = tI(,8 I a) * II(a) and appropriate independence assumptions, just like how the Bayesians derive their local computation scheme.
1302.6821
The Automated Mapping of Plans for Plan Recognition
cs.AI
To coordinate with other agents in its environment, an agent needs models of what the other agents are trying to do. When communication is impossible or expensive, this information must be acquired indirectly via plan recognition. Typical approaches to plan recognition start with a specification of the possible plans the other agents may be following, and develop special techniques for discriminating among the possibilities. Perhaps more desirable would be a uniform procedure for mapping plans to general structures supporting inference based on uncertain and incomplete observations. In this paper, we describe a set of methods for converting plans represented in a flexible procedural language to observation models represented as probabilistic belief networks.
1302.6822
A Logic for Default Reasoning About Probabilities
cs.AI
A logic is defined that allows to express information about statistical probabilities and about degrees of belief in specific propositions. By interpreting the two types of probabilities in one common probability space, the semantics given are well suited to model the influence of statistical information on the formation of subjective beliefs. Cross entropy minimization is a key element in these semantics, the use of which is justified by showing that the resulting logic exhibits some very reasonable properties.
1302.6823
Optimal Junction Trees
cs.AI
The paper deals with optimality issues in connection with updating beliefs in networks. We address two processes: triangulation and construction of junction trees. In the first part, we give a simple algorithm for constructing an optimal junction tree from a triangulated network. In the second part, we argue that any exact method based on local calculations must either be less efficient than the junction tree method, or it has an optimality problem equivalent to that of triangulation.
1302.6824
From Influence Diagrams to Junction Trees
cs.AI
We present an approach to the solution of decision problems formulated as influence diagrams. This approach involves a special triangulation of the underlying graph, the construction of a junction tree with special properties, and a message passing algorithm operating on the junction tree for computation of expected utilities and optimal decision policies.
1302.6825
Reduction of Computational Complexity in Bayesian Networks through Removal of Weak Dependencies
cs.AI
The paper presents a method for reducing the computational complexity of Bayesian networks through identification and removal of weak dependencies (removal of links from the (moralized) independence graph). The removal of a small number of links may reduce the computational complexity dramatically, since several fill-ins and moral links may be rendered superfluous by the removal. The method is described in terms of impact on the independence graph, the junction tree, and the potential functions associated with these. An empirical evaluation of the method using large real-world networks demonstrates the applicability of the method. Further, the method, which has been implemented in Hugin, complements the approximation method suggested by Jensen & Andersen (1990).
1302.6826
Using New Data to Refine a Bayesian Network
cs.AI
We explore the issue of refining an existent Bayesian network structure using new data which might mention only a subset of the variables. Most previous works have only considered the refinement of the network's conditional probability parameters, and have not addressed the issue of refining the network's structure. We develop a new approach for refining the network's structure. Our approach is based on the Minimal Description Length (MDL) principle, and it employs an adapted version of a Bayesian network learning algorithm developed in our previous work. One of the adaptations required is to modify the previous algorithm to account for the structure of the existent network. The learning algorithm generates a partial network structure which can then be used to improve the existent network. We also present experimental evidence demonstrating the effectiveness of our approach.
1302.6827
Syntax-based Default Reasoning as Probabilistic Model-based Diagnosis
cs.AI
We view the syntax-based approaches to default reasoning as a model-based diagnosis problem, where each source giving a piece of information is considered as a component. It is formalized in the ATMS framework (each source corresponds to an assumption). We assume then that all sources are independent and "fail" with a very small probability. This leads to a probability assignment on the set of candidates, or equivalently on the set of consistent environments. This probability assignment induces a Dempster-Shafer belief function which measures the probability that a proposition can be deduced from the evidence. This belief function can be used in several different ways to define a non-monotonic consequence relation. We study and compare these consequence relations. The -case of prioritized knowledge bases is briefly considered.
1302.6828
Induction of Selective Bayesian Classifiers
cs.LG stat.ML
In this paper, we examine previous work on the naive Bayesian classifier and review its limitations, which include a sensitivity to correlated features. We respond to this problem by embedding the naive Bayesian induction scheme within an algorithm that c arries out a greedy search through the space of features. We hypothesize that this approach will improve asymptotic accuracy in domains that involve correlated features without reducing the rate of learning in ones that do not. We report experimental results on six natural domains, including comparisons with decision-tree induction, that support these hypotheses. In closing, we discuss other approaches to extending naive Bayesian classifiers and outline some directions for future research.
1302.6829
Fuzzy Geometric Relations to Represent Hierarchical Spatial Information
cs.AI
A model to represent spatial information is presented in this paper. It is based on fuzzy constraints represented as fuzzy geometric relations that can be hierarchically structured. The concept of spatial template is introduced to capture the idea of interrelated objects in two-dimensional space. The representation model is used to specify imprecise or vague information consisting in relative locations and orientations of template objects. It is shown in this paper how a template represented by this model can be matched against a crisp situation to recognize a particular instance of this template. Furthermore, the proximity measure (fuzzy measure) between the instance and the template is worked out - this measure can be interpreted as a degree of similarity. In this context, template recognition can be viewed as a case of fuzzy pattern recognition. The results of this work have been implemented and applied to a complex military problem from which this work originated.
1302.6830
Constructing Belief Networks to Evaluate Plans
cs.AI
This paper examines the problem of constructing belief networks to evaluate plans produced by an knowledge-based planner. Techniques are presented for handling various types of complicating plan features. These include plans with context-dependent consequences, indirect consequences, actions with preconditions that must be true during the execution of an action, contingencies, multiple levels of abstraction multiple execution agents with partially-ordered and temporally overlapping actions, and plans which reference specific times and time durations.
1302.6831
Operator Selection While Planning Under Uncertainty
cs.AI
This paper describes the best first search strategy used by U-Plan (Mansell 1993a), a planning system that constructs quantitatively ranked plans given an incomplete description of an uncertain environment. U-Plan uses uncertain and incomplete evidence de scribing the environment, characterizes it using a Dempster-Shafer interval, and generates a set of possible world states. Plan construction takes place in an abstraction hierarchy where strategic decisions are made before tactical decisions. Search through this abstraction hierarchy is guided by a quantitative measure (expected fulfillment) based on decision theory. The search strategy is best first with the provision to update expected fulfillment and review previous decisions in the light of planning developments. U-Plan generates multiple plans for multiple possible worlds, and attempts to use existing plans for new world situations. A super-plan is then constructed, based on merging the set of plans and appropriately timed knowledge acquisition operators, which are used to decide between plan alternatives during plan execution.
1302.6832
Model-Based Diagnosis with Qualitative Temporal Uncertainty
cs.AI
In this paper we describe a framework for model-based diagnosis of dynamic systems, which extends previous work in this field by using and expressing temporal uncertainty in the form of qualitative interval relations a la Allen. Based on a logical framework extended by qualitative and quantitative temporal constraints we show how to describe behavioral models (both consistency- and abductive-based), discuss how to use abstract observations and show how abstract temporal diagnoses are computed. This yields an expressive framework, which allows the representation of complex temporal behavior allowing us to represent temporal uncertainty. Due to its abstraction capabilities computation is made independent of the number of observations and time points in a temporal setting. An example of hepatitis diagnosis is used throughout the paper.
1302.6833
Incremental Dynamic Construction of Layered Polytree Networks
cs.AI
Certain classes of problems, including perceptual data understanding, robotics, discovery, and learning, can be represented as incremental, dynamically constructed belief networks. These automatically constructed networks can be dynamically extended and modified as evidence of new individuals becomes available. The main result of this paper is the incremental extension of the singly connected polytree network in such a way that the network retains its singly connected polytree structure after the changes. The algorithm is deterministic and is guaranteed to have a complexity of single node addition that is at most of order proportional to the number of nodes (or size) of the network. Additional speed-up can be achieved by maintaining the path information. Despite its incremental and dynamic nature, the algorithm can also be used for probabilistic inference in belief networks in a fashion similar to other exact inference algorithms.
1302.6834
Models of Consensus for Multiple Agent Systems
cs.MA cs.SY
Models of consensus are used to manage multiple agent systems in order to choose between different recommendations provided by the system. It is assumed that there is a central agent that solicits recommendations or plans from other agents. That agent the n determines the consensus of the other agents, and chooses the resultant consensus recommendation or plan. Voting schemes such as this have been used in a variety of domains, including air traffic control. This paper uses an analytic model to study the use of consensus in multiple agent systems. The binomial model is used to study the probability that the consensus judgment is correct or incorrect. That basic model is extended to account for both different levels of agent competence and unequal prior odds. The analysis of that model is critical in the investigation of multiple agent systems, since the model leads us to conclude that in some cases consensus judgment is not appropriate. In addition, the results allow us to determine how many agents should be used to develop consensus decisions, which agents should be used to develop consensus decisions and under which conditions the consensus model should be used.
1302.6835
A Probabilistic Calculus of Actions
cs.AI
We present a symbolic machinery that admits both probabilistic and causal information about a given domain and produces probabilistic statements about the effect of actions and the impact of observations. The calculus admits two types of conditioning operators: ordinary Bayes conditioning, P(y|X = x), which represents the observation X = x, and causal conditioning, P(y|do(X = x)), read the probability of Y = y conditioned on holding X constant (at x) by deliberate action. Given a mixture of such observational and causal sentences, together with the topology of the causal graph, the calculus derives new conditional probabilities of both types, thus enabling one to quantify the effects of actions (and policies) from partially specified knowledge bases, such as Bayesian networks in which some conditional probabilities may not be available.
1302.6836
Robust Planning in Uncertain Environments
cs.AI
This paper describes a novel approach to planning which takes advantage of decision theory to greatly improve robustness in an uncertain environment. We present an algorithm which computes conditional plans of maximum expected utility. This algorithm relies on a representation of the search space as an AND/OR tree and employs a depth-limit to control computation costs. A numeric robustness factor, which parameterizes the utility function, allows the user to modulate the degree of risk-aversion employed by the planner. Via a look-ahead search, the planning algorithm seeks to find an optimal plan using expected utility as its optimization criterion. We present experimental results obtained by applying our algorithm to a non-deterministic extension of the blocks world domain. Our results demonstrate that the robustness factor governs the degree of risk embodied in the conditional plans computed by our algorithm.
1302.6837
Anytime Decision Making with Imprecise Probabilities
cs.AI
This paper examines methods of decision making that are able to accommodate limitations on both the form in which uncertainty pertaining to a decision problem can be realistically represented and the amount of computing time available before a decision must be made. The methods are anytime algorithms in the sense of Boddy and Dean 1991. Techniques are presented for use with Frisch and Haddawy's [1992] anytime deduction system, with an anytime adaptation of Nilsson's [1986] probabilistic logic, and with a probabilistic database model.
1302.6838
Three Approaches to Probability Model Selection
stat.ME cs.AI
This paper compares three approaches to the problem of selecting among probability models to fit data (1) use of statistical criteria such as Akaike's information criterion and Schwarz's "Bayesian information criterion," (2) maximization of the posterior probability of the model, and (3) maximization of an effectiveness ratio? trading off accuracy and computational cost. The unifying characteristic of the approaches is that all can be viewed as maximizing a penalized likelihood function. The second approach with suitable prior distributions has been shown to reduce to the first. This paper shows that the third approach reduces to the second for a particular form of the effectiveness ratio, and illustrates all three approaches with the problem of selecting the number of components in a mixture of Gaussian distributions. Unlike the first two approaches, the third can be used even when the candidate models are chosen for computational efficiency, without regard to physical interpretation, so that the likelihood and the prior distribution over models cannot be interpreted literally. As the most general and computationally oriented of the approaches, it is especially useful for artificial intelligence applications.
1302.6839
Knowledge Engineering for Large Belief Networks
cs.AI
We present several techniques for knowledge engineering of large belief networks (BNs) based on the our experiences with a network derived from a large medical knowledge base. The noisyMAX, a generalization of the noisy-OR gate, is used to model causal in dependence in a BN with multi-valued variables. We describe the use of leak probabilities to enforce the closed-world assumption in our model. We present Netview, a visualization tool based on causal independence and the use of leak probabilities. The Netview software allows knowledge engineers to dynamically view sub-networks for knowledge engineering, and it provides version control for editing a BN. Netview generates sub-networks in which leak probabilities are dynamically updated to reflect the missing portions of the network.
1302.6840
Solving Asymmetric Decision Problems with Influence Diagrams
cs.AI
While influence diagrams have many advantages as a representation framework for Bayesian decision problems, they have a serious drawback in handling asymmetric decision problems. To be represented in an influence diagram, an asymmetric decision problem must be symmetrized. A considerable amount of unnecessary computation may be involved when a symmetrized influence diagram is evaluated by conventional algorithms. In this paper we present an approach for avoiding such unnecessary computation in influence diagram evaluation.
1302.6841
Belief Maintenance in Bayesian Networks
cs.AI
Bayesian Belief Networks (BBNs) are a powerful formalism for reasoning under uncertainty but bear some severe limitations: they require a large amount of information before any reasoning process can start, they have limited contradiction handling capabilities, and their ability to provide explanations for their conclusion is still controversial. There exists a class of reasoning systems, called Truth Maintenance Systems (TMSs), which are able to deal with partially specified knowledge, to provide well-founded explanation for their conclusions, and to detect and handle contradictions. TMSs incorporating measure of uncertainty are called Belief Maintenance Systems (BMSs). This paper describes how a BMS based on probabilistic logic can be applied to BBNs, thus introducing a new class of BBNs, called Ignorant Belief Networks, able to incrementally deal with partially specified conditional dependencies, to provide explanations, and to detect and handle contradictions.
1302.6842
Belief Updating by Enumerating High-Probability Independence-Based Assignments
cs.AI
Independence-based (IB) assignments to Bayesian belief networks were originally proposed as abductive explanations. IB assignments assign fewer variables in abductive explanations than do schemes assigning values to all evidentially supported variables. We use IB assignments to approximate marginal probabilities in Bayesian belief networks. Recent work in belief updating for Bayes networks attempts to approximate posterior probabilities by finding a small number of the highest probability complete (or perhaps evidentially supported) assignments. Under certain assumptions, the probability mass in the union of these assignments is sufficient to obtain a good approximation. Such methods are especially useful for highly-connected networks, where the maximum clique size or the cutset size make the standard algorithms intractable. Since IB assignments contain fewer assigned variables, the probability mass in each assignment is greater than in the respective complete assignment. Thus, fewer IB assignments are sufficient, and a good approximation can be obtained more efficiently. IB assignments can be used for efficiently approximating posterior node probabilities even in cases which do not obey the rather strict skewness assumptions used in previous research. Two algorithms for finding the high probability IB assignments are suggested: one by doing a best-first heuristic search, and another by special-purpose integer linear programming. Experimental results show that this approach is feasible for highly connected belief networks.
1302.6843
Global Conditioning for Probabilistic Inference in Belief Networks
cs.AI
In this paper we propose a new approach to probabilistic inference on belief networks, global conditioning, which is a simple generalization of Pearl's (1986b) method of loopcutset conditioning. We show that global conditioning, as well as loop-cutset conditioning, can be thought of as a special case of the method of Lauritzen and Spiegelhalter (1988) as refined by Jensen et al (199Oa; 1990b). Nonetheless, this approach provides new opportunities for parallel processing and, in the case of sequential processing, a tradeoff of time for memory. We also show how a hybrid method (Suermondt and others 1990) combining loop-cutset conditioning with Jensen's method can be viewed within our framework. By exploring the relationships between these methods, we develop a unifying framework in which the advantages of each approach can be combined successfully.
1302.6844
Belief Induced by the Partial Knowledge of the Probabilities
cs.AI
We construct the belief function that quantifies the agent, beliefs about which event of Q will occurred when he knows that the event is selected by a chance set-up and that the probability function associated to the chance set up is only partially known.
1302.6845
Ignorance and the Expressiveness of Single- and Set-Valued Probability Models of Belief
cs.AI
Over time, there have hen refinements in the way that probability distributions are used for representing beliefs. Models which rely on single probability distributions depict a complete ordering among the propositions of interest, yet human beliefs are sometimes not completely ordered. Non-singleton sets of probability distributions can represent partially ordered beliefs. Convex sets are particularly convenient and expressive, but it is known that there are reasonable patterns of belief whose faithful representation require less restrictive sets. The present paper shows that prior ignorance about three or more exclusive alternatives and the emergence of partially ordered beliefs when evidence is obtained defy representation by any single set of distributions, but yield to a representation baud on several uts. The partial order is shown to be a partial qualitative probability which shares some intuitively appealing attributes with probability distributions.
1302.6846
A Probabilistic Approach to Hierarchical Model-based Diagnosis
cs.AI
Model-based diagnosis reasons backwards from a functional schematic of a system to isolate faults given observations of anomalous behavior. We develop a fully probabilistic approach to model based diagnosis and extend it to support hierarchical models. Our scheme translates the functional schematic into a Bayesian network and diagnostic inference takes place in the Bayesian network. A Bayesian network diagnostic inference algorithm is modified to take advantage of the hierarchy to give computational gains.
1302.6847
Semigraphoids Are Two-Antecedental Approximations of Stochastic Conditional Independence Models
cs.AI
The semigraphoid closure of every couple of CI-statements (GI=conditional independence) is a stochastic CI-model. As a consequence of this result it is shown that every probabilistically sound inference rule for CI-model, having at most two antecedents, is derivable from the semigraphoid inference rules. This justifies the use of semigraphoids as approximations of stochastic CI-models in probabilistic reasoning. The list of all 19 potential dominant elements of the mentioned semigraphoid closure is given as a byproduct.
1302.6848
Exceptional Subclasses in Qualitative Probability
cs.AI
System Z+ [Goldszmidt and Pearl, 1991, Goldszmidt, 1992] is a formalism for reasoning with normality defaults of the form "typically if phi then + (with strength cf)" where 6 is a positive integer. The system has a critical shortcoming in that it does not sanction inheritance across exceptional subclasses. In this paper we propose an extension to System Z+ that rectifies this shortcoming by extracting additional conditions between worlds from the defaults database. We show that the additional constraints do not change the notion of the consistency of a database. We also make comparisons with competing default reasoning systems.
1302.6849
A Defect in Dempster-Shafer Theory
cs.AI
By analyzing the relationships among chance, weight of evidence and degree of beliefwe show that the assertion "probability functions are special cases of belief functions" and the assertion "Dempster's rule can be used to combine belief functions based on distinct bodies of evidence" together lead to an inconsistency in Dempster-Shafer theory. To solve this problem, we must reject some fundamental postulates of the theory. We introduce a new approach for uncertainty management that shares many intuitive ideas with D-S theory, while avoiding this problem.
1302.6850
State-space Abstraction for Anytime Evaluation of Probabilistic Networks
cs.AI
One important factor determining the computational complexity of evaluating a probabilistic network is the cardinality of the state spaces of the nodes. By varying the granularity of the state spaces, one can trade off accuracy in the result for computational efficiency. We present an anytime procedure for approximate evaluation of probabilistic networks based on this idea. On application to some simple networks, the procedure exhibits a smooth improvement in approximation quality as computation time increases. This suggests that state-space abstraction is one more useful control parameter for designing real-time probabilistic reasoners.
1302.6851
General Belief Measures
cs.AI
Probability measures by themselves, are known to be inappropriate for modeling the dynamics of plain belief and their excessively strong measurability constraints make them unsuitable for some representational tasks, e.g. in the context of firstorder knowledge. In this paper, we are therefore going to look for possible alternatives and extensions. We begin by delimiting the general area of interest, proposing a minimal list of assumptions to be satisfied by any reasonable quasi-probabilistic valuation concept. Within this framework, we investigate two particularly interesting kinds of quasi-measures which are not or much less affected by the traditional problems. * Ranking measures, which generalize Spohn-type and possibility measures. * Cumulative measures, which combine the probabilistic and the ranking philosophy, allowing thereby a fine-grained account of static and dynamic belief.
1302.6852
Generating Graphoids from Generalised Conditional Probability
cs.AI
We take a general approach to uncertainty on product spaces, and give sufficient conditions for the independence structures of uncertainty measures to satisfy graphoid properties. Since these conditions are arguably more intuitive than some of the graphoid properties, they can be viewed as explanations why probability and certain other formalisms generate graphoids. The conditions include a sufficient condition for the Intersection property which can still apply even if there is a strong logical relations hip between the variables. We indicate how these results can be used to produce theories of qualitative conditional probability which are semi-graphoids and graphoids.
1302.6853
On Axiomatization of Probabilistic Conditional Independencies
cs.AI
This paper studies the connection between probabilistic conditional independence in uncertain reasoning and data dependency in relational databases. As a demonstration of the usefulness of this preliminary investigation, an alternate proof is presented for refuting the conjecture suggested by Pearl and Paz that probabilistic conditional independencies have a complete axiomatization.
1302.6854
Evidential Reasoning with Conditional Belief Functions
cs.AI
In the existing evidential networks with belief functions, the relations among the variables are always represented by joint belief functions on the product space of the involved variables. In this paper, we use conditional belief functions to represent such relations in the network and show some relations of these two kinds of representations. We also present a propagation algorithm for such networks. By analyzing the properties of some special evidential networks with conditional belief functions, we show that the reasoning process can be simplified in such kinds of networks.
1302.6855
Inter-causal Independence and Heterogeneous Factorization
cs.AI
It is well known that conditional independence can be used to factorize a joint probability into a multiplication of conditional probabilities. This paper proposes a constructive definition of inter-causal independence, which can be used to further factorize a conditional probability. An inference algorithm is developed, which makes use of both conditional independence and inter-causal independence to reduce inference complexity in Bayesian networks.
1302.6866
Vandermonde-subspace Frequency Division Multiplexing for Two-Tiered Cognitive Radio Networks
cs.IT math.IT
Vandermonde-subspace frequency division multiplexing (VFDM) is an overlay spectrum sharing technique for cognitive radio. VFDM makes use of a precoder based on a Vandermonde structure to transmit information over a secondary system, while keeping an orthogonal frequency division multiplexing (OFDM)-based primary system interference-free. To do so, VFDM exploits frequency selectivity and the use of cyclic prefixes by the primary system. Herein, a global view of VFDM is presented, including also practical aspects such as linear receivers and the impact of channel estimation. We show that VFDM provides a spectral efficiency increase of up to 1 bps/Hz over cognitive radio systems based on unused band detection. We also present some key design parameters for its future implementation and a feasible channel estimation protocol. Finally we show that, even when some of the theoretical assumptions are relaxed, VFDM provides non-negligible rates while protecting the primary system.
1302.6906
Tradition and Innovation in Scientists' Research Strategies
physics.soc-ph cs.DL cs.SI stat.AP
What factors affect a scientist's choice of research problem? Qualitative research in the history, philosophy, and sociology of science suggests that this choice is shaped by an "essential tension" between the professional demand for productivity and a conflicting drive toward risky innovation. We examine this tension empirically in the context of biomedical chemistry. We use complex networks to represent the evolving state of scientific knowledge, as expressed in publications. We then define research strategies relative to these networks. Scientists can introduce novel chemicals or chemical relationships--or delve deeper into known ones. They can consolidate existing knowledge clusters, or bridge distant ones. Analyzing such choices in aggregate, we find that the distribution of strategies remains remarkably stable, even as chemical knowledge grows dramatically. High-risk strategies, which explore new chemical relationships, are less prevalent in the literature, reflecting a growing focus on established knowledge at the expense of new opportunities. Research following a risky strategy is more likely to be ignored but also more likely to achieve high impact and recognition. While the outcome of a risky strategy has a higher expected reward than the outcome of a conservative strategy, the additional reward is insufficient to compensate for the additional risk. By studying the winners of 137 different prizes in biomedicine and chemistry, we show that the occasional "gamble" for extraordinary impact is the most plausible explanation for observed levels of risk-taking. Our empirical demonstration and unpacking of the "essential tension" suggests policy interventions that may foster more innovative research.
1302.6927
Online Learning for Time Series Prediction
cs.LG
In this paper we address the problem of predicting a time series using the ARMA (autoregressive moving average) model, under minimal assumptions on the noise terms. Using regret minimization techniques, we develop effective online learning algorithms for the prediction problem, without assuming that the noise terms are Gaussian, identically distributed or even independent. Furthermore, we show that our algorithm's performances asymptotically approaches the performance of the best ARMA model in hindsight.
1302.6932
Describing the complexity of systems: multi-variable "set complexity" and the information basis of systems biology
cs.IT math.IT q-bio.QM
Context dependence is central to the description of complexity. Keying on the pairwise definition of "set complexity" we use an information theory approach to formulate general measures of systems complexity. We examine the properties of multi-variable dependency starting with the concept of interaction information. We then present a new measure for unbiased detection of multi-variable dependency, "differential interaction information." This quantity for two variables reduces to the pairwise "set complexity" previously proposed as a context-dependent measure of information in biological systems. We generalize it here to an arbitrary number of variables. Critical limiting properties of the "differential interaction information" are key to the generalization. This measure extends previous ideas about biological information and provides a more sophisticated basis for study of complexity. The properties of "differential interaction information" also suggest new approaches to data analysis. Given a data set of system measurements differential interaction information can provide a measure of collective dependence, which can be represented in hypergraphs describing complex system interaction patterns. We investigate this kind of analysis using simulated data sets. The conjoining of a generalized set complexity measure, multi-variable dependency analysis, and hypergraphs is our central result. While our focus is on complex biological systems, our results are applicable to any complex system.
1302.6934
Optimum Header Positioning in Successive Interference Cancellation (SIC) based Aloha
cs.IT math.IT
Random Access MAC protocols are simple and effective when the nature of the traffic is unpredictable and sporadic. In the following paper, investigations on the new Enhanced Contention Resolution ALOHA (ECRA) are presented, where some new aspects of the protocol are investigated. Mathematical derivation and numerical evaluation of the symbol interference probability after SIC are here provided. Results of the optimum header positioning which is found to be in the beginning and in the end of the packets, are exploited for the evaluation of ECRA throughput and Packet Error Rate (PER) under imperfect knowledge of packets positions. Remarkable gains in the maximum throughput are observed for ECRA w.r.t. Contention Resolution ALOHA (CRA) under this assumption.
1302.6937
Online Convex Optimization Against Adversaries with Memory and Application to Statistical Arbitrage
cs.LG
The framework of online learning with memory naturally captures learning problems with temporal constraints, and was previously studied for the experts setting. In this work we extend the notion of learning with memory to the general Online Convex Optimization (OCO) framework, and present two algorithms that attain low regret. The first algorithm applies to Lipschitz continuous loss functions, obtaining optimal regret bounds for both convex and strongly convex losses. The second algorithm attains the optimal regret bounds and applies more broadly to convex losses without requiring Lipschitz continuity, yet is more complicated to implement. We complement our theoretic results with an application to statistical arbitrage in finance: we devise algorithms for constructing mean-reverting portfolios.
1302.6957
Ensemble Sparse Models for Image Analysis
cs.CV
Sparse representations with learned dictionaries have been successful in several image analysis applications. In this paper, we propose and analyze the framework of ensemble sparse models, and demonstrate their utility in image restoration and unsupervised clustering. The proposed ensemble model approximates the data as a linear combination of approximations from multiple \textit{weak} sparse models. Theoretical analysis of the ensemble model reveals that even in the worst-case, the ensemble can perform better than any of its constituent individual models. The dictionaries corresponding to the individual sparse models are obtained using either random example selection or boosted approaches. Boosted approaches learn one dictionary per round such that the dictionary learned in a particular round is optimized for the training examples having high reconstruction error in the previous round. Results with compressed recovery show that the ensemble representations lead to a better performance compared to using a single dictionary obtained with the conventional alternating minimization approach. The proposed ensemble models are also used for single image superresolution, and we show that they perform comparably to the recent approaches. In unsupervised clustering, experiments show that the proposed model performs better than baseline approaches in several standard datasets.
1302.6974
Spectrum Bandit Optimization
cs.LG cs.NI math.OC
We consider the problem of allocating radio channels to links in a wireless network. Links interact through interference, modelled as a conflict graph (i.e., two interfering links cannot be simultaneously active on the same channel). We aim at identifying the channel allocation maximizing the total network throughput over a finite time horizon. Should we know the average radio conditions on each channel and on each link, an optimal allocation would be obtained by solving an Integer Linear Program (ILP). When radio conditions are unknown a priori, we look for a sequential channel allocation policy that converges to the optimal allocation while minimizing on the way the throughput loss or {\it regret} due to the need for exploring sub-optimal allocations. We formulate this problem as a generic linear bandit problem, and analyze it first in a stochastic setting where radio conditions are driven by a stationary stochastic process, and then in an adversarial setting where radio conditions can evolve arbitrarily. We provide new algorithms in both settings and derive upper bounds on their regrets.
1302.6990
Stabilizer information inequalities from phase space distributions
quant-ph cs.IT math-ph math.IT math.MP
The Shannon entropy of a collection of random variables is subject to a number of constraints, the best-known examples being monotonicity and strong subadditivity. It remains an open question to decide which of these "laws of information theory" are also respected by the von Neumann entropy of many-body quantum states. In this article, we consider a toy version of this difficult problem by analyzing the von Neumann entropy of stabilizer states. We find that the von Neumann entropy of stabilizer states satisfies all balanced information inequalities that hold in the classical case. Our argument is built on the fact that stabilizer states have a classical model, provided by the discrete Wigner function: The phase-space entropy of the Wigner function corresponds directly to the von Neumann entropy of the state, which allows us to reduce to the classical case. Our result has a natural counterpart for multi-mode Gaussian states, which sheds some light on the general properties of the construction. We also discuss the relation of our results to recent work by Linden, Ruskai, and Winter.
1302.7025
Maximizing Acceptance Probability for Active Friending in On-Line Social Networks
cs.SI cs.CY physics.soc-ph
Friending recommendation has successfully contributed to the explosive growth of on-line social networks. Most friending recommendation services today aim to support passive friending, where a user passively selects friending targets from the recommended candidates. In this paper, we advocate recommendation support for active friending, where a user actively specifies a friending target. To the best of our knowledge, a recommendation designed to provide guidance for a user to systematically approach his friending target, has not been explored in existing on-line social networking services. To maximize the probability that the friending target would accept an invitation from the user, we formulate a new optimization problem, namely, \emph{Acceptance Probability Maximization (APM)}, and develop a polynomial time algorithm, called \emph{Selective Invitation with Tree and In-Node Aggregation (SITINA)}, to find the optimal solution. We implement an active friending service with SITINA in Facebook to validate our idea. Our user study and experimental results manifest that SITINA outperforms manual selection and the baseline approach in solution quality efficiently.
1302.7039
Content Based Image Retrieval System Using NOHIS-tree
cs.IR cs.CV cs.DB
Content-based image retrieval (CBIR) has been one of the most important research areas in computer vision. It is a widely used method for searching images in huge databases. In this paper we present a CBIR system called NOHIS-Search. The system is based on the indexing technique NOHIS-tree. The two phases of the system are described and the performance of the system is illustrated with the image database ImagEval. NOHIS-Search system was compared to other two CBIR systems; the first that using PDDP indexing algorithm and the second system is that using the sequential search. Results show that NOHIS-Search system outperforms the two other systems.
1302.7043
Scoup-SMT: Scalable Coupled Sparse Matrix-Tensor Factorization
stat.ML cs.LG
How can we correlate neural activity in the human brain as it responds to words, with behavioral data expressed as answers to questions about these same words? In short, we want to find latent variables, that explain both the brain activity, as well as the behavioral responses. We show that this is an instance of the Coupled Matrix-Tensor Factorization (CMTF) problem. We propose Scoup-SMT, a novel, fast, and parallel algorithm that solves the CMTF problem and produces a sparse latent low-rank subspace of the data. In our experiments, we find that Scoup-SMT is 50-100 times faster than a state-of-the-art algorithm for CMTF, along with a 5 fold increase in sparsity. Moreover, we extend Scoup-SMT to handle missing data without degradation of performance. We apply Scoup-SMT to BrainQ, a dataset consisting of a (nouns, brain voxels, human subjects) tensor and a (nouns, properties) matrix, with coupling along the nouns dimension. Scoup-SMT is able to find meaningful latent variables, as well as to predict brain activity with competitive accuracy. Finally, we demonstrate the generality of Scoup-SMT, by applying it on a Facebook dataset (users, friends, wall-postings); there, Scoup-SMT spots spammer-like anomalies.
1302.7051
Polyploidy and Discontinuous Heredity Effect on Evolutionary Multi-Objective Optimization
cs.NE
This paper examines the effect of mimicking discontinuous heredity caused by carrying more than one chromosome in some living organisms cells in Evolutionary Multi-Objective Optimization algorithms. In this representation, the phenotype may not fully reflect the genotype. By doing so we are mimicking living organisms inheritance mechanism, where traits may be silently carried for many generations to reappear later. Representations with different number of chromosomes in each solution vector are tested on different benchmark problems with high number of decision variables and objectives. A comparison with Non-Dominated Sorting Genetic Algorithm-II is done on all problems.
1302.7056
KSU KDD: Word Sense Induction by Clustering in Topic Space
cs.CL cs.AI stat.AP stat.ML
We describe our language-independent unsupervised word sense induction system. This system only uses topic features to cluster different word senses in their global context topic space. Using unlabeled data, this system trains a latent Dirichlet allocation (LDA) topic model then uses it to infer the topics distribution of the test instances. By clustering these topics distributions in their topic space we cluster them into different senses. Our hypothesis is that closeness in topic space reflects similarity between different word senses. This system participated in SemEval-2 word sense induction and disambiguation task and achieved the second highest V-measure score among all other systems.