id
stringlengths
9
16
title
stringlengths
4
278
categories
stringlengths
5
104
abstract
stringlengths
6
4.09k
1302.4916
Stacking from Tags: Clustering Bookmarks around a Theme
cs.IR
Since very recently, users on the social bookmarking service Delicious can stack web pages in addition to tagging them. Stacking enables users to group web pages around specific themes with the aim of recommending to others. However, users still stack a small subset of what they tag, and thus many web pages remain unstacked. This paper presents early research towards automatically clustering web pages from tags to find stacks and extend recommendations.
1302.4922
Structure Discovery in Nonparametric Regression through Compositional Kernel Search
stat.ML cs.LG stat.ME
Despite its importance, choosing the structural form of the kernel in nonparametric regression remains a black art. We define a space of kernel structures which are built compositionally by adding and multiplying a small number of base kernels. We present a method for searching over this space of structures which mirrors the scientific discovery process. The learned structures can often decompose functions into interpretable components and enable long-range extrapolation on time-series datasets. Our structure search method outperforms many widely used kernels and kernel combination methods on a variety of prediction tasks.
1302.4928
Graphical Models for Preference and Utility
cs.AI
Probabilistic independence can dramatically simplify the task of eliciting, representing, and computing with probabilities in large domains. A key technique in achieving these benefits is the idea of graphical modeling. We survey existing notions of independence for utility functions in a multi-attribute space, and suggest that these can be used to achieve similar advantages. Our new results concern conditional additive independence, which we show always has a perfect representation as separation in an undirected graph (a Markov network). Conditional additive independencies entail a particular functional for the utility function that is analogous to a product decomposition of a probability function, and confers analogous benefits. This functional form has been utilized in the Bayesian network and influence diagram literature, but generally without an explanation in terms of independence. The functional form yields a decomposition of the utility function that can greatly speed up expected utility calculations, particularly when the utility graph has a similar topology to the probabilistic network being used.
1302.4929
Counterfactuals and Policy Analysis in Structural Models
cs.AI
Evaluation of counterfactual queries (e.g., "If A were true, would C have been true?") is important to fault diagnosis, planning, determination of liability, and policy analysis. We present a method of revaluating counterfactuals when the underlying causal model is represented by structural models - a nonlinear generalization of the simultaneous equations models commonly used in econometrics and social sciences. This new method provides a coherent means for evaluating policies involving the control of variables which, prior to enacting the policy were influenced by other variables in the system.
1302.4930
Belief Functions and Default Reasoning
cs.AI
We present a new approach to dealing with default information based on the theory of belief functions. Our semantic structures, inspired by Adams' epsilon-semantics, are epsilon-belief assignments, where values committed to focal elements are either close to 0 or close to 1. We define two systems based on these structures, and relate them to other non-monotonic systems presented in the literature. We show that our second system correctly addresses the well-known problems of specificity, irrelevance, blocking of inheritance, ambiguity, and redundancy.
1302.4931
An Algebraic Semantics for Possibilistic Logic
cs.AI cs.LO
The first contribution of this paper is the presentation of a Pavelka - like formulation of possibilistic logic in which the language is naturally enriched by two connectives which represent negation (eg) and a new type of conjunction (otimes). The space of truth values for this logic is the lattice of possibility functions, that, from an algebraic point of view, forms a quantal. A second contribution comes from the understanding of the new conjunction as the combination of tokens of information coming from different sources, which makes our language "dynamic". A Gentzen calculus is presented, which is proved sound and complete with respect to the given semantics. The problem of truth functionality is discussed in this context.
1302.4932
Automating Computer Bottleneck Detection with Belief Nets
cs.AI
We describe an application of belief networks to the diagnosis of bottlenecks in computer systems. The technique relies on a high-level functional model of the interaction between application workloads, the Windows NT operating system, and system hardware. Given a workload description, the model predicts the values of observable system counters available from the Windows NT performance monitoring tool. Uncertainty in workloads, predictions, and counter values are characterized with Gaussian distributions. During diagnostic inference, we use observed performance monitor values to find the most probable assignment to the workload parameters. In this paper we provide some background on automated bottleneck detection, describe the structure of the system model, and discuss empirical procedures for model calibration and verification. Part of the calibration process includes generating a dataset to estimate a multivariate Gaussian error model. Initial results in diagnosing bottlenecks are presented.
1302.4933
Chain Graphs for Learning
cs.AI
Chain graphs combine directed and undirected graphs and their underlying mathematics combines properties of the two. This paper gives a simplified definition of chain graphs based on a hierarchical combination of Bayesian (directed) and Markov (undirected) networks. Examples of a chain graph are multivariate feed-forward networks, clustering with conditional interaction between variables, and forms of Bayes classifiers. Chain graphs are then extended using the notation of plates so that samples and data analysis problems can be represented in a graphical model as well. Implications for learning are discussed in the conclusion.
1302.4934
Error Estimation in Approximate Bayesian Belief Network Inference
cs.AI
We can perform inference in Bayesian belief networks by enumerating instantiations with high probability thus approximating the marginals. In this paper, we present a method for determining the fraction of instantiations that has to be considered such that the absolute error in the marginals does not exceed a predefined value. The method is based on extreme value theory. Essentially, the proposed method uses the reversed generalized Pareto distribution to model probabilities of instantiations below a given threshold. Based on this distribution, an estimate of the maximal absolute error if instantiations with probability smaller than u are disregarded can be made.
1302.4935
Generating the Structure of a Fuzzy Rule under Uncertainty
cs.AI
The aim of this paper is to present a method for identifying the structure of a rule in a fuzzy model. For this purpose, an ATMS shall be used (Zurita 1994). An algorithm obtaining the identification of the structure will be suggested (Castro 1995). The minimal structure of the rule (with respect to the number of variables that must appear in the rule) will be found by this algorithm. Furthermore, the identification parameters shall be obtained simultaneously. The proposed method shall be applied for classification to an example. The {em Iris Plant Database} shall be learnt for all three kinds of plants.
1302.4936
Practical Model-Based Diagnosis with Qualitative Possibilistic Uncertainty
cs.AI
An approach to fault isolation that exploits vastly incomplete models is presented. It relies on separate descriptions of each component behavior, together with the links between them, which enables focusing of the reasoning to the relevant part of the system. As normal observations do not need explanation, the behavior of the components is limited to anomaly propagation. Diagnostic solutions are disorders (fault modes or abnormal signatures) that are consistent with the observations, as well as abductive explanations. An ordinal representation of uncertainty based on possibility theory provides a simple exception-tolerant description of the component behaviors. We can for instance distinguish between effects that are more or less certainly present (or absent) and effects that are more or less certainly present (or absent) when a given anomaly is present. A realistic example illustrates the benefits of this approach.
1302.4937
Decision Flexibility
cs.AI
The development of new methods and representations for temporal decision-making requires a principled basis for characterizing and measuring the flexibility of decision strategies in the face of uncertainty. Our goal in this paper is to provide a framework - not a theory - for observing how decision policies behave in the face of informational perturbations, to gain clues as to how they might behave in the face of unanticipated, possibly unarticulated uncertainties. To this end, we find it beneficial to distinguish between two types of uncertainty: "Small World" and "Large World" uncertainty. The first type can be resolved by posing an unambiguous question to a "clairvoyant," and is anchored on some well-defined aspect of a decision frame. The second type is more troublesome, yet it is often of greater interest when we address the issue of flexibility; this type of uncertainty can be resolved only by consulting a "psychic." We next observe that one approach to flexibility used in the economics literature is already implicitly accounted for in the Maximum Expected Utility (MEU) principle from decision theory. Though simple, the observation establishes the context for a more illuminating notion of flexibility, what we term flexibility with respect to information revelation. We show how to perform flexibility analysis of a static (i.e., single period) decision problem using a simple example, and we observe that the most flexible alternative thus identified is not necessarily the MEU alternative. We extend our analysis for a dynamic (i.e., multi-period) model, and we demonstrate how to calculate the value of flexibility for decision strategies that allow downstream revision of an upstream commitment decision.
1302.4938
A Transformational Characterization of Equivalent Bayesian Network Structures
cs.AI
We present a simple characterization of equivalent Bayesian network structures based on local transformations. The significance of the characterization is twofold. First, we are able to easily prove several new invariant properties of theoretical interest for equivalent structures. Second, we use the characterization to derive an efficient algorithm that identifies all of the compelled edges in a structure. Compelled edge identification is of particular importance for learning Bayesian network structures from data because these edges indicate causal relationships when certain assumptions hold.
1302.4939
Conditioning Methods for Exact and Approximate Inference in Causal Networks
cs.AI
We present two algorithms for exact and approximate inference in causal networks. The first algorithm, dynamic conditioning, is a refinement of cutset conditioning that has linear complexity on some networks for which cutset conditioning is exponential. The second algorithm, B-conditioning, is an algorithm for approximate inference that allows one to trade-off the quality of approximations with the computation time. We also present some experimental results illustrating the properties of the proposed algorithms.
1302.4940
Independence Concepts for Convex Sets of Probabilities
cs.AI
In this paper we study different concepts of independence for convex sets of probabilities. There will be two basic ideas for independence. The first is irrelevance. Two variables are independent when a change on the knowledge about one variable does not affect the other. The second one is factorization. Two variables are independent when the joint convex set of probabilities can be decomposed on the product of marginal convex sets. In the case of the Theory of Probability, these two starting points give rise to the same definition. In the case of convex sets of probabilities, the resulting concepts will be strongly related, but they will not be equivalent. As application of the concept of independence, we shall consider the problem of building a global convex set from marginal convex sets of probabilities.
1302.4941
Clustering Without (Thinking About) Triangulation
cs.AI
The undirected technique for evaluating belief networks [Jensen, et.al., 1990, Lauritzen and Spiegelhalter, 1988] requires clustering the nodes in the network into a junction tree. In the traditional view, the junction tree is constructed from the cliques of the moralized and triangulated belief network: triangulation is taken to be the primitive concept, the goal towards which any clustering algorithm (e.g. node elimination) is directed. In this paper, we present an alternative conception of clustering, in which clusters and the junction tree property play the role of primitives: given a graph (not a tree) of clusters which obey (a modified version of) the junction tree property, we transform this graph until we have obtained a tree. There are several advantages to this approach: it is much clearer and easier to understand, which is important for humans who are constructing belief networks; it admits a wider range of heuristics which may enable more efficient or superior clustering algorithms; and it serves as the natural basis for an incremental clustering scheme, which we describe.
1302.4942
Implementation of Continuous Bayesian Networks Using Sums of Weighted Gaussians
cs.AI
Bayesian networks provide a method of representing conditional independence between random variables and computing the probability distributions associated with these random variables. In this paper, we extend Bayesian network structures to compute probability density functions for continuous random variables. We make this extension by approximating prior and conditional densities using sums of weighted Gaussian distributions and then finding the propagation rules for updating the densities in terms of these weights. We present a simple example that illustrates the Bayesian network for continuous variables; this example shows the effect of the network structure and approximation errors on the computation of densities for variables in the network.
1302.4943
Elicitation of Probabilities for Belief Networks: Combining Qualitative and Quantitative Information
cs.AI
Although the usefulness of belief networks for reasoning under uncertainty is widely accepted, obtaining numerical probabilities that they require is still perceived a major obstacle. Often not enough statistical data is available to allow for reliable probability estimation. Available information may not be directly amenable for encoding in the network. Finally, domain experts may be reluctant to provide numerical probabilities. In this paper, we propose a method for elicitation of probabilities from a domain expert that is non-invasive and accommodates whatever probabilistic information the expert is willing to state. We express all available information, whether qualitative or quantitative in nature, in a canonical form consisting of (in) equalities expressing constraints on the hyperspace of possible joint probability distributions. We then use this canonical form to derive second-order probability distributions over the desired probabilities.
1302.4944
Numerical Representations of Acceptance
cs.AI
Accepting a proposition means that our confidence in this proposition is strictly greater than the confidence in its negation. This paper investigates the subclass of uncertainty measures, expressing confidence, that capture the idea of acceptance, what we call acceptance functions. Due to the monotonicity property of confidence measures, the acceptance of a proposition entails the acceptance of any of its logical consequences. In agreement with the idea that a belief set (in the sense of Gardenfors) must be closed under logical consequence, it is also required that the separate acceptance o two propositions entail the acceptance of their conjunction. Necessity (and possibility) measures agree with this view of acceptance while probability and belief functions generally do not. General properties of acceptance functions are estabilished. The motivation behind this work is the investigation of a setting for belief revision more general than the one proposed by Alchourron, Gardenfors and Makinson, in connection with the notion of conditioning.
1302.4945
Fraud/Uncollectible Debt Detection Using a Bayesian Network Based Learning System: A Rare Binary Outcome with Mixed Data Structures
cs.AI
The fraud/uncollectible debt problem in the telecommunications industry presents two technical challenges: the detection and the treatment of the account given the detection. In this paper, we focus on the first problem of detection using Bayesian network models, and we briefly discuss the application of a normative expert system for the treatment at the end. We apply Bayesian network models to the problem of fraud/uncollectible debt detection for telecommunication services. In addition to being quite successful at predicting rare event outcomes, it is able to handle a mixture of categorical and continuous data. We present a performance comparison using linear and non-linear discriminant analysis, classification and regression trees, and Bayesian network models
1302.4946
A Constraint Satisfaction Approach to Decision under Uncertainty
cs.AI
The Constraint Satisfaction Problem (CSP) framework offers a simple and sound basis for representing and solving simple decision problems, without uncertainty. This paper is devoted to an extension of the CSP framework enabling us to deal with some decisions problems under uncertainty. This extension relies on a differentiation between the agent-controllable decision variables and the uncontrollable parameters whose values depend on the occurrence of uncertain events. The uncertainty on the values of the parameters is assumed to be given under the form of a probability distribution. Two algorithms are given, for computing respectively decisions solving the problem with a maximal probability, and conditional decisions mapping the largest possible amount of possible cases to actual decisions.
1302.4947
Plausibility Measures: A User's Guide
cs.AI
We examine a new approach to modeling uncertainty based on plausibility measures, where a plausibility measure just associates with an event its plausibility, an element is some partially ordered set. This approach is easily seen to generalize other approaches to modeling uncertainty, such as probability measures, belief functions, and possibility measures. The lack of structure in a plausibility measure makes it easy for us to add structure on an "as needed" basis, letting us examine what is required to ensure that a plausibility measure has certain properties of interest. This gives us insight into the essential features of the properties in question, while allowing us to prove general results that apply to many approaches to reasoning about uncertainty. Plausibility measures have already proved useful in analyzing default reasoning. In this paper, we examine their "algebraic properties," analogues to the use of + and * in probability theory. An understanding of such properties will be essential if plausibility measures are to be used in practice as a representation tool.
1302.4948
Testing Identifiability of Causal Effects
cs.AI
This paper concerns the probabilistic evaluation of the effects of actions in the presence of unmeasured variables. We show that the identification of causal effect between a singleton variable X and a set of variables Y can be accomplished systematically, in time polynomial in the number of variables in the graph. When the causal effect is identifiable, a closed-form expression can be obtained for the probability that the action will achieve a specified goal, or a set of goals.
1302.4949
A Characterization of the Dirichlet Distribution with Application to Learning Bayesian Networks
cs.AI cs.LG
We provide a new characterization of the Dirichlet distribution. This characterization implies that under assumptions made by several previous authors for learning belief networks, a Dirichlet prior on the parameters is inevitable.
1302.4950
Fast Belief Update Using Order-of-Magnitude Probabilities
cs.AI
We present an algorithm, called Predict, for updating beliefs in causal networks quantified with order-of-magnitude probabilities. The algorithm takes advantage of both the structure and the quantification of the network and presents a polynomial asymptotic complexity. Predict exhibits a conservative behavior in that it is always sound but not always complete. We provide sufficient conditions for completeness and present algorithms for testing these conditions and for computing a complete set of plausible values. We propose Predict as an efficient method to estimate probabilistic values and illustrate its use in conjunction with two known algorithms for probabilistic inference. Finally, we describe an application of Predict to plan evaluation, present experimental results, and discuss issues regarding its use with conditional logics of belief, and in the characterization of irrelevance.
1302.4951
Transforming Prioritized Defaults and Specificity into Parallel Defaults
cs.AI
We show how to transform any set of prioritized propositional defaults into an equivalent set of parallel (i.e., unprioritized) defaults, in circumscription. We give an algorithm to implement the transform. We show how to use the transform algorithm as a generator of a whole family of inferencing algorithms for circumscription. The method is to employ the transform algorithm as a front end to any inferencing algorithm, e.g., one of the previously available, that handles the parallel (empty) case of prioritization. Our algorithms provide not just coverage of a new expressive class, but also alternatives to previous algorithms for implementing the previously covered class (?layered?) of prioritization. In particular, we give a new query-answering algorithm for prioritized cirumscription which is sound and complete for the full expressive class of unrestricted finite prioritization partial orders, for propositional defaults (or minimized predicates). By contrast, previous algorithms required that the prioritization partial order be layered, i.e., structured similar to the system of rank in the military. Our algorithm enables, for the first time, the implementation of the most useful class of prioritization: non-layered prioritization partial orders. Default inheritance, for example, typically requires non-layered prioritization to represent specificity adequately. Our algorithm enables not only the implementation of default inheritance (and specificity) within prioritized circumscription, but also the extension and combination of default inheritance with other kinds of prioritized default reasoning, e.g.: with stratified logic programs with negation-as-failure. Such logic programs are previously known to be representable equivalently as layered-priority predicate circumscriptions. Worst-case, the transform increases the number of defaults exponentially. We discuss how inferencing is practically implementable nevertheless in two kinds of situations: general expressiveness but small numbers of defaults, or expressive special cases with larger numbers of defaults. One such expressive special case is non-?top-heaviness? of the prioritization partial order. In addition to its direct implementation, the transform can also be exploited analytically to generate special case algorithms, e.g., a tractable transform for a class within default inheritance (detailed in another, forthcoming paper). We discuss other aspects of the significance of the fundamental result. One can view the transform as reducing n degrees of partially ordered belief confidence to just 2 degrees of confidence: for-sure and (unprioritized) default. Ordinary, parallel default reasoning, e.g., in parallel circumscription or Poole's Theorist, can be viewed in these terms as reducing 2 degrees of confidence to just 1 degree of confidence: that of the non-monotonic theory's conclusions. The expressive reduction's computational complexity suggests that prioritization is valuable for its expressive conciseness, just as defaults are for theirs. For Reiter's Default Logic and Poole's Theorist, the transform implies how to extend those formalisms so as to equip them with a concept of prioritization that is exactly equivalent to that in circumscription. This provides an interesting alternative to Brewka's approach to equipping them with prioritization-type precedence.
1302.4952
Efficient Decision-Theoretic Planning: Techniques and Empirical Analysis
cs.AI
This paper discusses techniques for performing efficient decision-theoretic planning. We give an overview of the DRIPS decision-theoretic refinement planning system, which uses abstraction to efficiently identify optimal plans. We present techniques for automatically generating search control information, which can significantly improve the planner's performance. We evaluate the efficiency of DRIPS both with and without the search control rules on a complex medical planning problem and compare its performance to that of a branch-and-bound decision tree algorithm.
1302.4953
Fuzzy Logic and Probability
cs.AI
In this paper we deal with a new approach to probabilistic reasoning in a logical framework. Nearly almost all logics of probability that have been proposed in the literature are based on classical two-valued logic. After making clear the differences between fuzzy logic and probability theory, here we propose a {em fuzzy} logic of probability for which completeness results (in a probabilistic sense) are provided. The main idea behind this approach is that probability values of crisp propositions can be understood as truth-values of some suitable fuzzy propositions associated to the crisp ones. Moreover, suggestions and examples of how to extend the formalism to cope with conditional probabilities and with other uncertainty formalisms are also provided.
1302.4954
Probabilistic Temporal Reasoning with Endogenous Change
cs.AI
This paper presents a probabilistic model for reasoning about the state of a system as it changes over time, both due to exogenous and endogenous influences. Our target domain is a class of medical prediction problems that are neither so urgent as to preclude careful diagnosis nor progress so slowly as to allow arbitrary testing and treatment options. In these domains there is typically enough time to gather information about the patient's state and consider alternative diagnoses and treatments, but the temporal interaction between the timing of tests, treatments, and the course of the disease must also be considered. Our approach is to elicit a qualitative structural model of the patient from a human expert---the model identifies important attributes, the way in which exogenous changes affect attribute values, and the way in which the patient's condition changes endogenously. We then elicit probabilistic information to capture the expert's uncertainty about the effects of tests and treatments and the nature and timing of endogenous state changes. This paper describes the model in the context of a problem in treating vehicle accident trauma, and suggests a method for solving the model based on the technique of sequential imputation. A complementary goal of this work is to understand and synthesize a disparate collection of research efforts all using the name ?probabilistic temporal reasoning.? This paper analyzes related work and points out essential differences between our proposed model and other approaches in the literature.
1302.4955
Toward a Characterization of Uncertainty Measure for the Dempster-Shafer Theory
cs.AI
This is a working paper summarizing results of an ongoing research project whose aim is to uniquely characterize the uncertainty measure for the Dempster-Shafer Theory. A set of intuitive axiomatic requirements is presented, some of their implications are shown, and the proof is given of the minimality of recently proposed measure AU among all measures satisfying the proposed requirements.
1302.4956
A Definition and Graphical Representation for Causality
cs.AI
We present a precise definition of cause and effect in terms of a fundamental notion called unresponsiveness. Our definition is based on Savage's (1954) formulation of decision theory and departs from the traditional view of causation in that our causal assertions are made relative to a set of decisions. An important consequence of this departure is that we can reason about cause locally, not requiring a causal explanation for every dependency. Such local reasoning can be beneficial because it may not be necessary to determine whether a particular dependency is causal to make a decision. Also in this paper, we examine the graphical encoding of causal relationships. We show that influence diagrams in canonical form are an accurate and efficient representation of causal relationships. In addition, we establish a correspondence between canonical form and Pearl's causal theory.
1302.4957
Learning Bayesian Networks: A Unification for Discrete and Gaussian Domains
cs.AI
We examine Bayesian methods for learning Bayesian networks from a combination of prior knowledge and statistical data. In particular, we unify the approaches we presented at last year's conference for discrete and Gaussian domains. We derive a general Bayesian scoring metric, appropriate for both domains. We then use this metric in combination with well-known statistical facts about the Dirichlet and normal--Wishart distributions to derive our metrics for discrete and Gaussian domains.
1302.4958
A Bayesian Approach to Learning Causal Networks
cs.AI
Whereas acausal Bayesian networks represent probabilistic independence, causal Bayesian networks represent causal relationships. In this paper, we examine Bayesian methods for learning both types of networks. Bayesian methods for learning acausal networks are fairly well developed. These methods often employ assumptions to facilitate the construction of priors, including the assumptions of parameter independence, parameter modularity, and likelihood equivalence. We show that although these assumptions also can be appropriate for learning causal networks, we need additional assumptions in order to learn causal networks. We introduce two sufficient assumptions, called {em mechanism independence} and {em component independence}. We show that these new assumptions, when combined with parameter independence, parameter modularity, and likelihood equivalence, allow us to apply methods for learning acausal networks to learn causal networks.
1302.4959
Display of Information for Time-Critical Decision Making
cs.AI
We describe methods for managing the complexity of information displayed to people responsible for making high-stakes, time-critical decisions. The techniques provide tools for real-time control of the configuration and quantity of information displayed to a user, and a methodology for designing flexible human-computer interfaces for monitoring applications. After defining a prototypical set of display decision problems, we introduce the expected value of revealed information (EVRI) and the related measure of expected value of displayed information (EVDI). We describe how these measures can be used to enhance computer displays used for monitoring complex systems. We motivate the presentation by discussing our efforts to employ decision-theoretic control of displays for a time-critical monitoring application at the NASA Mission Control Center in Houston.
1302.4960
Reasoning, Metareasoning, and Mathematical Truth: Studies of Theorem Proving under Limited Resources
cs.AI
In earlier work, we introduced flexible inference and decision-theoretic metareasoning to address the intractability of normative inference. Here, rather than pursuing the task of computing beliefs and actions with decision models composed of distinctions about uncertain events, we examine methods for inferring beliefs about mathematical truth before an automated theorem prover completes a proof. We employ a Bayesian analysis to update belief in truth, given theorem-proving progress, and show how decision-theoretic methods can be used to determine the value of continuing to deliberate versus taking immediate action in time-critical situations.
1302.4961
Improved Sampling for Diagnostic Reasoning in Bayesian Networks
cs.AI
Bayesian networks offer great potential for use in automating large scale diagnostic reasoning tasks. Gibbs sampling is the main technique used to perform diagnostic reasoning in large richly interconnected Bayesian networks. Unfortunately Gibbs sampling can take an excessive time to generate a representative sample. In this paper we describe and test a number of heuristic strategies for improving sampling in noisy-or Bayesian networks. The strategies include Monte Carlo Markov chain sampling techniques other than Gibbs sampling. Emphasis is put on strategies that can be implemented in distributed systems.
1302.4962
Cautious Propagation in Bayesian Networks
cs.AI
Consider the situation where some evidence e has been entered to a Bayesian network. When performing conflict analysis, sensitivity analysis, or when answering questions like "What if the finding on X had been y instead of x?" you need probabilities P (e'| h), where e' is a subset of e, and h is a configuration of a (possibly empty) set of variables. Cautious propagation is a modification of HUGIN propagation into a Shafer-Shenoy-like architecture. It is less efficient than HUGIN propagation; however, it provides easy access to P (e'| h) for a great deal of relevant subsets e'.
1302.4963
Information/Relevance Influence Diagrams
cs.AI
In this paper we extend the influence diagram (ID) representation for decisions under uncertainty. In the standard ID, arrows into a decision node are only informational; they do not represent constraints on what the decision maker can do. We can represent such constraints only indirectly, using arrows to the children of the decision and sometimes adding more variables to the influence diagram, thus making the ID more complicated. Users of influence diagrams often want to represent constraints by arrows into decision nodes. We represent constraints on decisions by allowing relevance arrows into decision nodes. We call the resulting representation information/relevance influence diagrams (IRIDs). Information/relevance influence diagrams allow for direct representation and specification of constrained decisions. We use a combination of stochastic dynamic programming and Gibbs sampling to solve IRIDs. This method is especially useful when exact methods for solving IDs fail.
1302.4964
Estimating Continuous Distributions in Bayesian Classifiers
cs.LG cs.AI stat.ML
When modeling a probability distribution with a Bayesian network, we are faced with the problem of how to handle continuous variables. Most previous work has either solved the problem by discretizing, or assumed that the data are generated by a single Gaussian. In this paper we abandon the normality assumption and instead use statistical methods for nonparametric density estimation. For a naive Bayesian classifier, we present experimental results on a variety of natural and artificial domains, comparing two methods of density estimation: assuming normality and modeling each conditional distribution with a single Gaussian; and using nonparametric kernel density estimation. We observe large reductions in error on several natural and artificial data sets, which suggests that kernel estimation is a useful tool for learning Bayesian models.
1302.4965
Stochastic Simulation Algorithms for Dynamic Probabilistic Networks
cs.AI
Stochastic simulation algorithms such as likelihood weighting often give fast, accurate approximations to posterior probabilities in probabilistic networks, and are the methods of choice for very large networks. Unfortunately, the special characteristics of dynamic probabilistic networks (DPNs), which are used to represent stochastic temporal processes, mean that standard simulation algorithms perform very poorly. In essence, the simulation trials diverge further and further from reality as the process is observed over time. In this paper, we present simulation algorithms that use the evidence observed at each time step to push the set of trials back towards reality. The first algorithm, "evidence reversal" (ER) restructures each time slice of the DPN so that the evidence nodes for the slice become ancestors of the state variables. The second algorithm, called "survival of the fittest" sampling (SOF), "repopulates" the set of trials at each time step using a stochastic reproduction rate weighted by the likelihood of the evidence according to each trial. We compare the performance of each algorithm with likelihood weighting on the original network, and also investigate the benefits of combining the ER and SOF methods. The ER/SOF combination appears to maintain bounded error independent of the number of time steps in the simulation.
1302.4966
Probabilistic Exploration in Planning while Learning
cs.AI
Sequential decision tasks with incomplete information are characterized by the exploration problem; namely the trade-off between further exploration for learning more about the environment and immediate exploitation of the accrued information for decision-making. Within artificial intelligence, there has been an increasing interest in studying planning-while-learning algorithms for these decision tasks. In this paper we focus on the exploration problem in reinforcement learning and Q-learning in particular. The existing exploration strategies for Q-learning are of a heuristic nature and they exhibit limited scaleability in tasks with large (or infinite) state and action spaces. Efficient experimentation is needed for resolving uncertainties when possible plans are compared (i.e. exploration). The experimentation should be sufficient for selecting with statistical significance a locally optimal plan (i.e. exploitation). For this purpose, we develop a probabilistic hill-climbing algorithm that uses a statistical selection procedure to decide how much exploration is needed for selecting a plan which is, with arbitrarily high probability, arbitrarily close to a locally optimal one. Due to its generality the algorithm can be employed for the exploration strategy of robust Q-learning. An experiment on a relatively complex control task shows that the proposed exploration strategy performs better than a typical exploration strategy.
1302.4967
On the Detection of Conflicts in Diagnostic Bayesian Networks Using Abstraction
cs.AI
An important issue in the use of expert systems is the so-called brittleness problem. Expert systems model only a limited part of the world. While the explicit management of uncertainty in expert systems itigates the brittleness problem, it is still possible for a system to be used, unwittingly, in ways that the system is not prepared to address. Such a situation may be detected by the method of straw models, first presented by Jensen et al. [1990] and later generalized and justified by Laskey [1991]. We describe an algorithm, which we have implemented, that takes as input an annotated diagnostic Bayesian network (the base model) and constructs, without assistance, a bipartite network to be used as a straw model. We show that in some cases this straw model is better that the independent straw model of Jensen et al., the only other straw model for which a construction algorithm has been designed and implemented.
1302.4968
HUGS: Combining Exact Inference and Gibbs Sampling in Junction Trees
cs.AI
Dawid, Kjaerulff and Lauritzen (1994) provided a preliminary description of a hybrid between Monte-Carlo sampling methods and exact local computations in junction trees. Utilizing the strengths of both methods, such hybrid inference methods has the potential of expanding the class of problems which can be solved under bounded resources as well as solving problems which otherwise resist exact solutions. The paper provides a detailed description of a particular instance of such a hybrid scheme; namely, combination of exact inference and Gibbs sampling in discrete Bayesian networks. We argue that this combination calls for an extension of the usual message passing scheme of ordinary junction trees.
1302.4969
Sensitivities: An Alternative to Conditional Probabilities for Bayesian Belief Networks
cs.AI
We show an alternative way of representing a Bayesian belief network by sensitivities and probability distributions. This representation is equivalent to the traditional representation by conditional probabilities, but makes dependencies between nodes apparent and intuitively easy to understand. We also propose a QR matrix representation for the sensitivities and/or conditional probabilities which is more efficient, in both memory requirements and computational speed, than the traditional representation for computer-based implementations of probabilistic inference. We use sensitivities to show that for a certain class of binary networks, the computation time for approximate probabilistic inference with any positive upper bound on the error of the result is independent of the size of the network. Finally, as an alternative to traditional algorithms that use conditional probabilities, we describe an exact algorithm for probabilistic inference that uses the QR-representation for sensitivities and updates probability distributions of nodes in a network according to messages from the neighbors.
1302.4970
Is There a Role for Qualitative Risk Assessment?
cs.AI
Classically, risk is characterized by a point value probability indicating the likelihood of occurrence of an adverse effect. However, there are domains where the attainability of objective numerical risk characterizations is increasingly being questioned. This paper reviews the arguments in favour of extending classical techniques of risk assessment to incorporate meaningful qualitative and weak quantitative risk characterizations. A technique in which linguistic uncertainty terms are defined in terms of patterns of argument is then proposed. The technique is demonstrated using a prototype computer-based system for predicting the carcinogenic risk due to novel chemical compounds.
1302.4971
On the Complexity of Solving Markov Decision Problems
cs.AI
Markov decision problems (MDPs) provide the foundations for a number of problems of interest to AI researchers studying automated planning and reinforcement learning. In this paper, we summarize results regarding the complexity of solving MDPs and the running time of MDP solution algorithms. We argue that, although MDPs can be solved efficiently in theory, more study is needed to reveal practical algorithms for solving large problems quickly. To encourage future research, we sketch some alternative methods of analysis that rely on the structure of MDPs.
1302.4972
Causal Inference and Causal Explanation with Background Knowledge
cs.AI
This paper presents correct algorithms for answering the following two questions; (i) Does there exist a causal explanation consistent with a set of background knowledge which explains all of the observed independence facts in a sample? (ii) Given that there is such a causal explanation what are the causal relationships common to every such causal explanation?
1302.4973
Strong Completeness and Faithfulness in Bayesian Networks
cs.AI
A completeness result for d-separation applied to discrete Bayesian networks is presented and it is shown that in a strong measure-theoretic sense almost all discrete distributions for a given network structure are faithful; i.e. the independence facts true of the distribution are all and only those entailed by the network structure.
1302.4974
A Theoretical Framework for Context-Sensitive Temporal Probability Model Construction with Application to Plan Projection
cs.AI
We define a context-sensitive temporal probability logic for representing classes of discrete-time temporal Bayesian networks. Context constraints allow inference to be focused on only the relevant portions of the probabilistic knowledge. We provide a declarative semantics for our language. We present a Bayesian network construction algorithm whose generated networks give sound and complete answers to queries. We use related concepts in logic programming to justify our approach. We have implemented a Bayesian network construction algorithm for a subset of the theory and demonstrate it's application to the problem of evaluating the effectiveness of treatments for acute cardiac conditions.
1302.4975
Refining Reasoning in Qualitative Probabilistic Networks
cs.AI
In recent years there has been a spate of papers describing systems for probabilisitic reasoning which do not use numerical probabilities. In some cases the simple set of values used by these systems make it impossible to predict how a probability will change or which hypothesis is most likely given certain evidence. This paper concentrates on such situations, and suggests a number of ways in which they may be resolved by refining the representation.
1302.4976
On the Testability of Causal Models with Latent and Instrumental Variables
cs.AI
Certain causal models involving unmeasured variables induce no independence constraints among the observed variables but imply, nevertheless, inequality contraints on the observed distribution. This paper derives a general formula for such instrumental variables, that is, exogenous variables that directly affect some variables but not all. With the help of this formula, it is possible to test whether a model involving instrumental variables may account for the data, or, conversely, whether a given variables can be deemed instrumental.
1302.4977
Probabilistic Evaluation of Sequential Plans from Causal Models with Hidden Variables
cs.AI
The paper concerns the probabilistic evaluation of plans in the presence of unmeasured variables, each plan consisting of several concurrent or sequential actions. We establish a graphical criterion for recognizing when the effects of a given plan can be predicted from passive observations on measured variables only. When the criterion is satisfied, a closed-form expression is provided for the probability that the plan will achieve a specified goal.
1302.4978
Exploiting the Rule Structure for Decision Making within the Independent Choice Logic
cs.AI
This paper introduces the independent choice logic, and in particular the "single agent with nature" instance of the independent choice logic, namely ICLdt. This is a logical framework for decision making uncertainty that extends both logic programming and stochastic models such as influence diagrams. This paper shows how the representation of a decision problem within the independent choice logic can be exploited to cut down the combinatorics of dynamic programming. One of the main problems with influence diagram evaluation techniques is the need to optimise a decision for all values of the 'parents' of a decision variable. In this paper we show how the rule based nature of the ICLdt can be exploited so that we only make distinctions in the values of the information available for a decision that will make a difference to utility.
1302.4979
Abstraction in Belief Networks: The Role of Intermediate States in Diagnostic Reasoning
cs.AI
Bayesian belief networks are bing increasingly used as a knowledge representation for diagnostic reasoning. One simple method for conducting diagnostic reasoning is to represent system faults and observations only. In this paper, we investigate how having intermediate nodes-nodes other than fault and observation nodes affects the diagnostic performance of a Bayesian belief network. We conducted a series of experiments on a set of real belief networks for medical diagnosis in liver and bile disease. We compared the effects on diagnostic performance of a two-level network consisting just of disease and finding nodes with that of a network which models intermediate pathophysiological disease states as well. We provide some theoretical evidence for differences observed between the abstracted two-level network and the full network.
1302.4980
Accounting for Context in Plan Recognition, with Application to Traffic Monitoring
cs.AI
Typical approaches to plan recognition start from a representation of an agent's possible plans, and reason evidentially from observations of the agent's actions to assess the plausibility of the various candidates. A more expansive view of the task (consistent with some prior work) accounts for the context in which the plan was generated, the mental state and planning process of the agent, and consequences of the agent's actions in the world. We present a general Bayesian framework encompassing this view, and focus on how context can be exploited in plan recognition. We demonstrate the approach on a problem in traffic monitoring, where the objective is to induce the plan of the driver from observation of vehicle movements. Starting from a model of how the driver generates plans, we show how the highway context can appropriately influence the recognizer's interpretation of observed driver behavior.
1302.4981
A New Pruning Method for Solving Decision Trees and Game Trees
cs.AI
The main goal of this paper is to describe a new pruning method for solving decision trees and game trees. The pruning method for decision trees suggests a slight variant of decision trees that we call scenario trees. In scenario trees, we do not need a conditional probability for each edge emanating from a chance node. Instead, we require a joint probability for each path from the root node to a leaf node. We compare the pruning method to the traditional rollback method for decision trees and game trees. For problems that require Bayesian revision of probabilities, a scenario tree representation with the pruning method is more efficient than a decision tree representation with the rollback method. For game trees, the pruning method is more efficient than the rollback method.
1302.4982
Directed Cyclic Graphical Representations of Feedback Models
cs.AI
The use of directed acyclic graphs (DAGs) to represent conditional independence relations among random variables has proved fruitful in a variety of ways. Recursive structural equation models are one kind of DAG model. However, non-recursive structural equation models of the kinds used to model economic processes are naturally represented by directed cyclic graphs with independent errors, a characterization of conditional independence errors, a characterization of conditional independence constraints is obtained, and it is shown that the result generalizes in a natural way to systems in which the error variables or noises are statistically dependent. For non-linear systems with independent errors a sufficient condition for conditional independence of variables in associated distributions is obtained.
1302.4983
Causal Inference in the Presence of Latent Variables and Selection Bias
cs.AI
We show that there is a general, informative and reliable procedure for discovering causal relations when, for all the investigator knows, both latent variables and selection bias may be at work. Given information about conditional independence and dependence relations between measured variables, even when latent variables and selection bias may be present, there are sufficient conditions for reliably concluding that there is a causal path from one variable to another, and sufficient conditions for reliably concluding when no such causal path exists.
1302.4984
Modeling Failure Priors and Persistence in Model-Based Diagnosis
cs.AI
Probabilistic model-based diagnosis computes the posterior probabilities of failure of components from the prior probabilities of component failure and observations of system behavior. One problem with this method is that such priors are almost never directly available. One of the reasons is that the prior probability estimates include an implicit notion of a time interval over which they are specified -- for example, if the probability of failure of a component is 0.05, is this over the period of a day or is this over a week? A second problem facing probabilistic model-based diagnosis is the modeling of persistence. Say we have an observation about a system at time t_1 and then another observation at a later time t_2. To compute posterior probabilities that take into account both the observations, we need some model of how the state of the system changes from time t_1 to t_2. In this paper, we address these problems using techniques from Reliability theory. We show how to compute the failure prior of a component from an empirical measure of its reliability -- the Mean Time Between Failure (MTBF). We also develop a scheme to model persistence when handling multiple time tagged observations.
1302.4985
A Polynomial Algorithm for Computing the Optimal Repair Strategy in a System with Independent Component Failures
cs.AI
The goal of diagnosis is to compute good repair strategies in response to anomalous system behavior. In a decision theoretic framework, a good repair strategy has low expected cost. In a general formulation of the problem, the computation of the optimal (lowest expected cost) repair strategy for a system with multiple faults is intractable. In this paper, we consider an interesting and natural restriction on the behavior of the system being diagnosed: (a) the system exhibits faulty behavior if and only if one or more components is malfunctioning. (b) The failures of the system components are independent. Given this restriction on system behavior, we develop a polynomial time algorithm for computing the optimal repair strategy. We then go on to introduce a system hierarchy and the notion of inspecting (testing) components before repair. We develop a linear time algorithm for computing an optimal repair strategy for the hierarchical system which includes both repair and inspection.
1302.4986
Exploiting System Hierarchy to Compute Repair Plans in Probabilistic Model-based Diagnosis
cs.AI
The goal of model-based diagnosis is to isolate causes of anomalous system behavior and recommend inexpensive repair actions in response. In general, precomputing optimal repair policies is intractable. To date, investigators addressing this problem have explored approximations that either impose restrictions on the system model (such as a single fault assumption) or compute an immediate best action with limited lookahead. In this paper, we develop a formulation of repair in model-based diagnosis and a repair algorithm that computes optimal sequences of actions. This optimal approach is costly but can be applied to precompute an optimal repair strategy for compact systems. We show how we can exploit a hierarchical system specification to make this approach tractable for large systems. When introducing hierarchy, we also consider the tradeoff between simply replacing a component and decomposing it to repair its subcomponents. The hierarchical repair algorithm is suitable for off-line precomputation of an optimal repair strategy. A modification of the algorithm takes advantage of an iterative deepening scheme to trade off inference time and the quality of the computed strategy.
1302.4987
Path Planning under Time-Dependent Uncertainty
cs.AI
Standard algorithms for finding the shortest path in a graph require that the cost of a path be additive in edge costs, and typically assume that costs are deterministic. We consider the problem of uncertain edge costs, with potential probabilistic dependencies among the costs. Although these dependencies violate the standard dynamic-programming decomposition, we identify a weaker stochastic consistency condition that justifies a generalized dynamic-programming approach based on stochastic dominance. We present a revised path-planning algorithm and prove that it produces optimal paths under time-dependent uncertain costs. We test the algorithm by applying it to a model of stochastic bus networks, and present empirical performance results comparing it to some alternatives. Finally, we consider extensions of these concepts to a more general class of problems of heuristic search under uncertainty.
1302.4988
Defaults and Infinitesimals: Defeasible Inference by Nonarchimedean Entropy-Maximization
cs.AI
We develop a new semantics for defeasible inference based on extended probability measures allowed to take infinitesimal values, on the interpretation of defaults as generalized conditional probability constraints and on a preferred-model implementation of entropy maximization.
1302.4989
An Order of Magnitude Calculus
cs.AI
This paper develops a simple calculus for order of magnitude reasoning. A semantics is given with soundness and completeness results. Order of magnitude probability functions are easily defined and turn out to be equivalent to kappa functions, which are slight generalizations of Spohn's Natural Conditional Functions. The calculus also gives rise to an order of magnitude decision theory, which can be used to justify an amended version of Pearl's decision theory for kappa functions, although the latter is weaker and less expressive.
1302.4990
A Method for Implementing a Probabilistic Model as a Relational Database
cs.AI
This paper discusses a method for implementing a probabilistic inference system based on an extended relational data model. This model provides a unified approach for a variety of applications such as dynamic programming, solving sparse linear equations, and constraint propagation. In this framework, the probability model is represented as a generalized relational database. Subsequent probabilistic requests can be processed as standard relational queries. Conventional database management systems can be easily adopted for implementing such an approximate reasoning system.
1302.4991
Optimization of Inter-Subnet Belief Updating in Multiply Sectioned Bayesian Networks
cs.AI
Recent developments show that Multiply Sectioned Bayesian Networks (MSBNs) can be used for diagnosis of natural systems as well as for model-based diagnosis of artificial systems. They can be applied to single-agent oriented reasoning systems as well as multi-agent distributed probabilistic reasoning systems. Belief propagation between a pair of subnets plays a central role in maintenance of global consistency in a MSBN. This paper studies the operation UpdateBelief, presented originally with MSBNs, for inter-subnet propagation. We analyze how the operation achieves its intended functionality, which provides hints as for how its efficiency can be improved. We then define two new versions of UpdateBelief that reduce the computation time for inter-subnet propagation. One of them is optimal in the sense that the minimum amount of computation for coordinating multi-linkage belief propagation is required. The optimization problem is solved through the solution of a graph-theoretic problem: the minimum weight open tour in a tree.
1302.4992
Generating Explanations for Evidential Reasoning
cs.AI
In this paper, we present two methods to provide explanations for reasoning with belief functions in the valuation-based systems. One approach, inspired by Strat's method, is based on sensitivity analysis, but its computation is simpler thus easier to implement than Strat's. The other one is to examine the impact of evidence on the conclusion based on the measure of the information content in the evidence. We show the property of additivity for the pieces of evidence that are conditional independent within the context of the valuation-based systems. We will give an example to show how these approaches are applied in an evidential network.
1302.4993
Inference with Causal Independence in the CPSC Network
cs.AI
This paper reports experiments with the causal independence inference algorithm proposed by Zhang and Poole (1994b) on the CPSC network created by Pradhan et al. (1994). It is found that the algorithm is able to answer 420 of the 422 possible zero-observation queries, 94 of 100 randomly generated five-observation queries, 87 of 100 randomly generated ten-observation queries, and 69 of 100 randomly generated twenty-observation queries.
1302.5002
Asymptotic Data Rates of Receive-Diversity Systems with MMSE Estimation and Spatially Correlated Interferers
cs.IT math.IT
An asymptotic technique is presented to characterize the bits/symbol achievable on a representative wireless link in a spatially distributed network with active interferers at correlated positions, N receive diversity branches, and linear Minimum-Mean-Square-Error (MMSE) receivers. This framework is then applied to systems including analogs to Matern type I and type II networks which are useful to model systems with Medium-Access Control (MAC), cellular uplinks with orthogonal transmissions and frequency reuse, and Boolean cluster networks. It is found that for our network models, with moderately large N, the correlation between interferer positions does not significantly influence the bits/symbol resulting in simple approximations for the data rates achievable in such networks which are known to be difficult to analyze and for which few analytical results are available.
1302.5010
Matching Pursuit LASSO Part II: Applications and Sparse Recovery over Batch Signals
cs.CV cs.LG stat.ML
Matching Pursuit LASSIn Part I \cite{TanPMLPart1}, a Matching Pursuit LASSO ({MPL}) algorithm has been presented for solving large-scale sparse recovery (SR) problems. In this paper, we present a subspace search to further improve the performance of MPL, and then continue to address another major challenge of SR -- batch SR with many signals, a consideration which is absent from most of previous $\ell_1$-norm methods. As a result, a batch-mode {MPL} is developed to vastly speed up sparse recovery of many signals simultaneously. Comprehensive numerical experiments on compressive sensing and face recognition tasks demonstrate the superior performance of MPL and BMPL over other methods considered in this paper, in terms of sparse recovery ability and efficiency. In particular, BMPL is up to 400 times faster than existing $\ell_1$-norm methods considered to be state-of-the-art.O Part II: Applications and Sparse Recovery over Batch Signals
1302.5021
Linear Coding Schemes for the Distributed Computation of Subspaces
cs.IT math.IT
Let $X_1, ..., X_m$ be a set of $m$ statistically dependent sources over the common alphabet $\mathbb{F}_q$, that are linearly independent when considered as functions over the sample space. We consider a distributed function computation setting in which the receiver is interested in the lossless computation of the elements of an $s$-dimensional subspace $W$ spanned by the elements of the row vector $[X_1, \ldots, X_m]\Gamma$ in which the $(m \times s)$ matrix $\Gamma$ has rank $s$. A sequence of three increasingly refined approaches is presented, all based on linear encoders. The first approach uses a common matrix to encode all the sources and a Korner-Marton like receiver to directly compute $W$. The second improves upon the first by showing that it is often more efficient to compute a carefully chosen superspace $U$ of $W$. The superspace is identified by showing that the joint distribution of the $\{X_i\}$ induces a unique decomposition of the set of all linear combinations of the $\{X_i\}$, into a chain of subspaces identified by a normalized measure of entropy. This subspace chain also suggests a third approach, one that employs nested codes. For any joint distribution of the $\{X_i\}$ and any $W$, the sum-rate of the nested code approach is no larger than that under the Slepian-Wolf (SW) approach. Under the SW approach, $W$ is computed by first recovering each of the $\{X_i\}$. For a large class of joint distributions and subspaces $W$, the nested code approach is shown to improve upon SW. Additionally, a class of source distributions and subspaces are identified, for which the nested-code approach is sum-rate optimal.
1302.5039
Cognitive Interference Alignment for OFDM Two-tiered Networks
cs.IT math.IT
In this contribution, we introduce an interference alignment scheme that allows the coexistence of an orthogonal frequency division multiplexing (OFDM) macro-cell and a cognitive small-cell, deployed in a two-tiered structure and transmitting over the same bandwidth. We derive the optimal linear strategy for the single antenna secondary base station, maximizing the spectral efficiency of the opportunistic link, accounting for both signal sub-space structure and power loading strategy. Our analytical and numerical findings prove that the precoder structure proposed is optimal for the considered scenario in the face of Rayleigh and exponential decaying channels.
1302.5056
Pooling-Invariant Image Feature Learning
cs.CV cs.LG
Unsupervised dictionary learning has been a key component in state-of-the-art computer vision recognition architectures. While highly effective methods exist for patch-based dictionary learning, these methods may learn redundant features after the pooling stage in a given early vision architecture. In this paper, we offer a novel dictionary learning scheme to efficiently take into account the invariance of learned features after the spatial pooling stage. The algorithm is built on simple clustering, and thus enjoys efficiency and scalability. We discuss the underlying mechanism that justifies the use of clustering algorithms, and empirically show that the algorithm finds better dictionaries than patch-based methods with the same dictionary size.
1302.5082
Proceedings of the Third International Workshop on Domain-Specific Languages and Models for Robotic Systems (DSLRob 2012)
cs.RO
Proceedings of the Third International Workshop on Domain-Specific Languages and Models for Robotic Systems (DSLRob'12), held at the 2012 International Conference on Simulation, Modeling, and Programming for Autonomous Robots (SIMPAR 2012), November 2012 in Tsukuba, Japan. The main topics of the workshop were Domain-Specific Languages (DSLs) and Model-driven Architecture (MDA) for robotics. A domain-specific language (DSL) is a programming language dedicated to a particular problem domain that offers specific notations and abstractions that increase programmer productivity within that domain. Models-driven architecture (MDA) offers a high-level way for domain users to specify the functionality of their system at the right level of abstraction. DSLs and models have historically been used for programming complex systems. However recently they have garnered interest as a separate field of study. Robotic systems blend hardware and software in a holistic way that intrinsically raises many crosscutting concerns (concurrency, uncertainty, time constraints, ...), for which reason, traditional general-purpose languages often lead to a poor fit between the language features and the implementation requirements. DSLs and models offer a powerful, systematic way to overcome this problem, enabling the programmer to quickly and precisely implement novel software solutions to complex problems within the robotics domain.
1302.5085
Model-driven engineering approach to design and implementation of robot control system
cs.RO cs.SE
In this paper we apply a model-driven engineering approach to designing domain-specific solutions for robot control system development. We present a case study of the complete process, including identification of the domain meta-model, graphical notation definition and source code generation for subsumption architecture -- a well-known example of robot control architecture. Our goal is to show that both the definition of the robot-control architecture and its supporting tools fits well into the typical workflow of model-driven engineering development.
1302.5125
High-Dimensional Probability Estimation with Deep Density Models
stat.ML cs.LG
One of the fundamental problems in machine learning is the estimation of a probability distribution from data. Many techniques have been proposed to study the structure of data, most often building around the assumption that observations lie on a lower-dimensional manifold of high probability. It has been more difficult, however, to exploit this insight to build explicit, tractable density models for high-dimensional data. In this paper, we introduce the deep density model (DDM), a new approach to density estimation. We exploit insights from deep learning to construct a bijective map to a representation space, under which the transformation of the distribution of the data is approximately factorized and has identical and known marginal densities. The simplicity of the latent distribution under the model allows us to feasibly explore it, and the invertibility of the map to characterize contraction of measure across it. This enables us to compute normalized densities for out-of-sample data. This combination of tractability and flexibility allows us to tackle a variety of probabilistic tasks on high-dimensional datasets, including: rapid computation of normalized densities at test-time without evaluating a partition function; generation of samples without MCMC; and characterization of the joint entropy of the data.
1302.5130
Quantum-inspired Huffman Coding
cs.IT cs.ET math.IT
Huffman Compression, also known as Huffman Coding, is one of many compression techniques in use today. The two important features of Huffman coding are instantaneousness that is the codes can be interpreted as soon as they are received and variable length that is a most frequent symbol has length smaller than a less frequent symbol. The traditional Huffman coding has two procedures: constructing a tree in O(n^2) and then traversing it in O(n). Quantum computing is a promising approach of computation that is based on equations from Quantum Mechanics. Instantaneousness and variable length features are difficult to generalize to the quantum case. The quantum coding field is pioneered by Schumacher works on block coding scheme. To encode N signals sequentially, it requires O(N3) computational steps. The encoding and decoding processes are far from instantaneous. Moreover, the lengths of all the codewords are the same. A Huffman-coding-inspired scheme for the storage of quantum information takes O(N(log N)a) computational steps for a sequential implementation on non-parallel machines.
1302.5145
Prediction and Clustering in Signed Networks: A Local to Global Perspective
cs.SI cs.LG
The study of social networks is a burgeoning research area. However, most existing work deals with networks that simply encode whether relationships exist or not. In contrast, relationships in signed networks can be positive ("like", "trust") or negative ("dislike", "distrust"). The theory of social balance shows that signed networks tend to conform to some local patterns that, in turn, induce certain global characteristics. In this paper, we exploit both local as well as global aspects of social balance theory for two fundamental problems in the analysis of signed networks: sign prediction and clustering. Motivated by local patterns of social balance, we first propose two families of sign prediction methods: measures of social imbalance (MOIs), and supervised learning using high order cycles (HOCs). These methods predict signs of edges based on triangles and \ell-cycles for relatively small values of \ell. Interestingly, by examining measures of social imbalance, we show that the classic Katz measure, which is used widely in unsigned link prediction, actually has a balance theoretic interpretation when applied to signed networks. Furthermore, motivated by the global structure of balanced networks, we propose an effective low rank modeling approach for both sign prediction and clustering. For the low rank modeling approach, we provide theoretical performance guarantees via convex relaxations, scale it up to large problem sizes using a matrix factorization based algorithm, and provide extensive experimental validation including comparisons with local approaches. Our experimental results indicate that, by adopting a more global viewpoint of balance structure, we get significant performance and computational gains in prediction and clustering tasks on signed networks. Our work therefore highlights the usefulness of the global aspect of balance theory for the analysis of signed networks.
1302.5150
Measuring Agglomeration of Agglomerated Particles Pictures
cs.CE math-ph math.AT math.MP math.NA math.PR
In this article, we introduce a novel geometrical index $\delta_{agg}$, which is associated with the Euler number and is obtained by an image processing procedure for a given digital picture of aggregated particles such that $\delta_{agg}$ exhibits the degree of the agglomerations of the particles. In the previous work (Matsutani, Shimosako, Wang, Appl.Math.Modeling {\bf{37}} (2013), 4007-4022), we proposed an algorithm to construct a picture of agglomerated particles as a Monte-Carlo simulation whose agglomeration degree is controlled by $\gamma_{agg} \in (0,1)$. By applying the image processing procedure to the pictures of the agglomeration particles constructed following the algorithm, we show that $\delta_{agg}$ statistically reproduces the agglomeration parameter $\gamma_{agg}$.
1302.5153
Constructing Polar Codes Using Iterative Bit-Channel Upgrading
cs.IT math.IT
The definition of polar codes given by Arikan is explicit, but the construction complexity is an issue. This is due to the exponential growth in the size of the output alphabet of the bit-channels as the codeword length increases. Tal and Vardy recently presented a method for constructing polar codes which controls this growth. They approximated each bit-channel with a better channel and a worse channel while reducing the alphabet size. They constructed a polar code based on the worse channel and used the better channel to measure the distance from the optimal channel. This paper considers the knowledge gained from the perspective of the better channel. A method is presented using iterative upgrading of the bit-channels which successively results in a channel closer to the original one. It is shown that this approach can be used to obtain a channel arbitrarily close to the original channel, and therefore to the optimal construction of a polar code.
1302.5166
Rate-Compatible Short-Length Protograph LDPC Codes
cs.IT math.IT
This paper produces a rate-compatible protograph LDPC code at 1k information blocklength with superior performance in both waterfall and error floor regions. The design of such codes has proved difficult in the past because the constraints imposed by structured design (protographs), rate-compatibility, as well as small block length, are not easily satisfied together. For example, as the block length decreases, the predominance of decoding threshold as the main parameter in coding design is reduced, thus complicating the search for good codes. Our rate-compatible protograph codes have rates ranging from 1/3 to 4/5 and show no error floor down to $10^{-6}$ FER.
1302.5168
q-ary Compressive Sensing
cs.IT math.IT math.ST stat.TH
We introduce q-ary compressive sensing, an extension of 1-bit compressive sensing. We propose a novel sensing mechanism and a corresponding recovery procedure. The recovery properties of the proposed approach are analyzed both theoretically and empirically. Results in 1-bit compressive sensing are recovered as a special case. Our theoretical results suggest a tradeoff between the quantization parameter q, and the number of measurements m in the control of the error of the resulting recovery algorithm, as well its robustness to noise.
1302.5181
Basic Classes of Grammars with Prohibition
cs.FL cs.CL
A practical tool for natural language modeling and development of human-machine interaction is developed in the context of formal grammars and languages. A new type of formal grammars, called grammars with prohibition, is introduced. Grammars with prohibition provide more powerful tools for natural language generation and better describe processes of language learning than the conventional formal grammars. Here we study relations between languages generated by different grammars with prohibition based on conventional types of formal grammars such as context-free or context sensitive grammars. Besides, we compare languages generated by different grammars with prohibition and languages generated by conventional formal grammars. In particular, it is demonstrated that they have essentially higher computational power and expressive possibilities in comparison with the conventional formal grammars. Thus, while conventional formal grammars are recursive and subrecursive algorithms, many classes of grammars with prohibition are superrecursive algorithms. Results presented in this work are aimed at the development of human-machine interaction, modeling natural languages, empowerment of programming languages, computer simulation, better software systems, and theory of recursion.
1302.5186
Unsupervised edge map scoring: a statistical complexity approach
cs.CV stat.AP
We propose a new Statistical Complexity Measure (SCM) to qualify edge maps without Ground Truth (GT) knowledge. The measure is the product of two indices, an \emph{Equilibrium} index $\mathcal{E}$ obtained by projecting the edge map into a family of edge patterns, and an \emph{Entropy} index $\mathcal{H}$, defined as a function of the Kolmogorov Smirnov (KS) statistic. This new measure can be used for performance characterization which includes: (i)~the specific evaluation of an algorithm (intra-technique process) in order to identify its best parameters, and (ii)~the comparison of different algorithms (inter-technique process) in order to classify them according to their quality. Results made over images of the South Florida and Berkeley databases show that our approach significantly improves over Pratt's Figure of Merit (PFoM) which is the objective reference-based edge map evaluation standard, as it takes into account more features in its evaluation.
1302.5189
Object Detection in Real Images
cs.CV
Object detection and recognition are important problems in computer vision. Since these problems are meta-heuristic, despite a lot of research, practically usable, intelligent, real-time, and dynamic object detection/recognition methods are still unavailable. We propose a new object detection/recognition method, which improves over the existing methods in every stage of the object detection/recognition process. In addition to the usual features, we propose to use geometric shapes, like linear cues, ellipses and quadrangles, as additional features. The full potential of geometric cues is exploited by using them to extract other features in a robust, computationally efficient, and less meta-heuristic manner. We also propose a new hierarchical codebook, which provides good generalization and discriminative properties. The codebook enables fast multi-path inference mechanisms based on propagation of conditional likelihoods, that make it robust to occlusion and noise. It has the capability of dynamic learning. We also propose a new learning method that has generative and discriminative learning capabilities, does not need large and fully supervised training dataset, and is capable of online learning. The preliminary work of detecting geometric shapes in real images has been completed. This preliminary work is the focus of this report. Future path for realizing the proposed object detection/recognition method is also discussed in brief.
1302.5205
The exponential family in abstract information theory
cs.IT math-ph math.IT math.MP
We introduce generalized notions of a divergence function and a Fisher information matrix. We propose to generalize the notion of an exponential family of models by reformulating it in terms of the Fisher information matrix. Our methods are those of information geometry. The context is general enough to include applications from outside statistics.
1302.5215
Development Of Ontology-Based Intelligent System For Software Testing
cs.AI cs.SE
Software testing is a prime factor in software industry. Besides knowing the importance of testing, only limited time is allocated for teaching it. It will be more efficient if testing is taught simultaneously with programming foundations. This integrated learning of testing techniques and programming allows the programmers to perform in a better way and this leads to the improvement of the performance of the industry progress. In this paper, a technique named ontology is introduced, it first defines the various testing process in hierarchy and define relationships among them, to share and reuse the knowledge that is captured, secondly metadata is created by natural language processing and finally, the application use ontologies to support test management, it act as knowledge base for multiple environment with the integrated teaching of programming foundation and testing concepts. Keywords: Meta Data, Ontology, Software Testing, Integration, Programming Foundations.
1302.5226
Dobrushin ergodicity coefficient for Markov operators on cones, and beyond
math.OA cs.MA math.DG
The analysis of classical consensus algorithms relies on contraction properties of adjoints of Markov operators, with respect to Hilbert's projective metric or to a related family of seminorms (Hopf's oscillation or Hilbert's seminorm). We generalize these properties to abstract consensus operators over normal cones, which include the unital completely positive maps (Kraus operators) arising in quantum information theory. In particular, we show that the contraction rate of such operators, with respect to the Hopf oscillation seminorm, is given by an analogue of Dobrushin's ergodicity coefficient. We derive from this result a characterization of the contraction rate of a non-linear flow, with respect to Hopf's oscillation seminorm and to Hilbert's projective metric.
1302.5235
Predicting the Temporal Dynamics of Information Diffusion in Social Networks
cs.SI physics.soc-ph
Online social networks play a major role in the spread of information at very large scale and it becomes essential to provide means to analyse this phenomenon. In this paper we address the issue of predicting the temporal dynamics of the information diffusion process. We develop a graph-based approach built on the assumption that the macroscopic dynamics of the spreading process are explained by the topology of the network and the interactions that occur through it, between pairs of users, on the basis of properties at the microscopic level. We introduce a generic model, called T-BaSIC, and describe how to estimate its parameters from users behaviours using machine learning techniques. Contrary to classical approaches where the parameters are fixed in advance, T-BaSIC's parameters are functions depending of time, which permit to better approximate and adapt to the diffusion phenomenon observed in online social networks. Our proposal has been validated on real Twitter datasets. Experiments show that our approach is able to capture the particular patterns of diffusion depending of the studied sub-networks of users and topics. The results corroborate the "two-step" theory (1955) that states that information flows from media to a few "opinion leaders" who then transfer it to the mass population via social networks and show that it applies in the online context. This work also highlights interesting recommendations for future investigations.
1302.5280
Opportunistic Interference Alignment for MIMO Interfering Multiple-Access Channels
cs.IT math.IT
We consider the $K$-cell multiple-input multiple-output (MIMO) interfering multiple-access channel (IMAC) with time-invariant channel coefficients, where each cell consists of a base station (BS) with $M$ antennas and $N$ users having $L$ antennas each. In this paper, we propose two opportunistic interference alignment (OIA) techniques utilizing multiple transmit antennas at each user: antenna selection-based OIA and singular value decomposition (SVD)-based OIA. Their performance is analyzed in terms of \textit{user scaling law} required to achieve $KS$ degrees-of-freedom (DoF), where $S(\le M)$ denotes the number of simultaneously transmitting users per cell. We assume that each selected user transmits a single data stream at each time-slot. It is shown that the antenna selection-based OIA does not fundamentally change the user scaling condition if $L$ is fixed, compared with the single-input multiple-output (SIMO) IMAC case, which is given by $\text{SNR}^{(K 1)S}$, where SNR denotes the signal-to-noise ratio. In addition, we show that the SVD-based OIA can greatly reduce the user scaling condition to $\text{SNR}^{(K-1)S-L+1}$ through optimizing a weight vector at each user. Simulation results validate the derived scaling laws of the proposed OIA techniques. The sum-rate performance of the proposed OIA techniques is compared with the conventional techniques in MIMO IMAC channels and it is shown that the proposed OIA techniques outperform the conventional techniques.
1302.5281
Fundamental bound on the reliability of quantum information transmission
quant-ph cs.IT math.IT
Information theory tells us that if the rate of sending information across a noisy channel were above the capacity of that channel, then the transmission would necessarily be unreliable. For classical information sent over classical or quantum channels, one could, under certain conditions, make a stronger statement that the reliability of the transmission shall decay exponentially to zero with the number of channel uses and the proof of this statement typically relies on a certain fundamental bound on the reliability of the transmission. Such a statement or the bound has never been given for sending quantum information. We give this bound and then use it to give the first example where the reliability of sending quantum information at rates above the capacity decays exponentially to zero. We also show that our framework can be used for proving generalized bounds on the reliability.
1302.5302
Dynamic Memory Allocation Policies for Postings in Real-Time Twitter Search
cs.IR cs.DB
We explore a real-time Twitter search application where tweets are arriving at a rate of several thousands per second. Real-time search demands that they be indexed and searchable immediately, which leads to a number of implementation challenges. In this paper, we focus on one aspect: dynamic postings allocation policies for index structures that are completely held in main memory. The core issue can be characterized as a "Goldilocks Problem". Because memory remains today a scare resource, an allocation policy that is too aggressive leads to inefficient utilization, while a policy that is too conservative is slow and leads to fragmented postings lists. We present a dynamic postings allocation policy that allocates memory in increasingly-larger "slices" from a small number of large, fixed pools of memory. Through analytical models and experiments, we explore different settings that balance time (query evaluation speed) and space (memory utilization).
1302.5348
Graph-based Generalization Bounds for Learning Binary Relations
cs.LG
We investigate the generalizability of learned binary relations: functions that map pairs of instances to a logical indicator. This problem has application in numerous areas of machine learning, such as ranking, entity resolution and link prediction. Our learning framework incorporates an example labeler that, given a sequence $X$ of $n$ instances and a desired training size $m$, subsamples $m$ pairs from $X \times X$ without replacement. The challenge in analyzing this learning scenario is that pairwise combinations of random variables are inherently dependent, which prevents us from using traditional learning-theoretic arguments. We present a unified, graph-based analysis, which allows us to analyze this dependence using well-known graph identities. We are then able to bound the generalization error of learned binary relations using Rademacher complexity and algorithmic stability. The rate of uniform convergence is partially determined by the labeler's subsampling process. We thus examine how various assumptions about subsampling affect generalization; under a natural random subsampling process, our bounds guarantee $\tilde{O}(1/\sqrt{n})$ uniform convergence.
1302.5371
Non-Linear Distributed Average Consensus using Bounded Transmissions
cs.DC cs.IT math.IT
A distributed average consensus algorithm in which every sensor transmits with bounded peak power is proposed. In the presence of communication noise, it is shown that the nodes reach consensus asymptotically to a finite random variable whose expectation is the desired sample average of the initial observations with a variance that depends on the step size of the algorithm and the variance of the communication noise. The asymptotic performance is characterized by deriving the asymptotic covariance matrix using results from stochastic approximation theory. It is shown that using bounded transmissions results in slower convergence compared to the linear consensus algorithm based on the Laplacian heuristic. Simulations corroborate our analytical findings.
1302.5374
A Weight-coded Evolutionary Algorithm for the Multidimensional Knapsack Problem
cs.NE math.OC
A revised weight-coded evolutionary algorithm (RWCEA) is proposed for solving multidimensional knapsack problems. This RWCEA uses a new decoding method and incorporates a heuristic method in initialization. Computational results show that the RWCEA performs better than a weight-coded evolutionary algorithm proposed by Raidl (1999) and to some existing benchmarks, it can yield better results than the ones reported in the OR-library.
1302.5376
Spatial CSIT Allocation Policies for Network MIMO Channels
cs.IT math.IT
In this work, we study the problem of the optimal dissemination of channel state information (CSI) among K spatially distributed transmitters (TXs) jointly cooperating to serve K receivers (RXs). One of the particularities of this work lies in the fact that the CSI is distributed in the sense that each TX obtains its own estimate of the global multi-user MIMO channel with no further exchange of information being allowed between the TXs. Although this is well suited to model the cooperation between non-colocated TXs, e.g., in cellular Coordinated Multipoint (CoMP) schemes, this type of setting has received little attention so far in the information theoretic society. We study in this work what are the CSI requirements at every TX, as a function of the network geometry, to ensure that the maximal number of degrees-of-freedom (DoF) is achieved, i.e., the same DoF as obtained under perfect CSI at all TXs. We advocate the use of the generalized DoF to take into account the geometry of the network in the analysis. Consistent with the intuition, the derived DoF maximizing CSI allocation policy suggests that TX cooperation should be limited to a specific finite neighborhood around each TX. This is in sharp contrast with the conventional (uniform) CSI dissemination policy which induces CSI requirements that grow unbounded with the network size. The proposed CSI allocation policy suggests an alternative to clustering which overcomes fundamental limitations such as (i) edge interference and (ii) unbounded increase of the CSIT requirements with the cluster size. Finally, we show how finite neighborhood CSIT exchange translates into finite neighborhood message exchange so that finally global interference management is possible with only local cooperation
1302.5383
Stochastic Ordering of Fading Channels Through the Shannon Transform
cs.IT math.IT
A new stochastic order between two fading distributions is introduced. A fading channel dominates another in the ergodic capacity ordering sense, if the Shannon transform of the first is greater than that of the second at all values of average signal to noise ratio. It is shown that some parametric fading models such as the Nakagami-m, Rician, and Hoyt are distributions that are monotonic in their line of sight parameters with respect to the ergodic capacity order. Some operations under which the ergodic capacity order is preserved are also discussed. Through these properties of the ergodic capacity order, it is possible to compare under two different fading scenarios, the ergodic capacity of a composite system involving multiple fading links with coding/decoding capabilities only at the transmitter/receiver. Such comparisons can be made even in cases when a closed form expression for the ergodic capacity of the composite system is not analytically tractable. Applications to multiple access channels, and extensions to multiple-input multiple-output (MIMO) systems are also discussed.
1302.5384
Facilitating Machine to Machine (M2M) Communication using GSM Network
cs.IT cs.NI math.IT
In this paper a method to facilitate M2M communication using existing GSM networks is proposed - as M2M devices primarily use SMS as their data bearer, the focus is on increasing the number of devices that can use the associated GSM signaling channels at a time. This is achieved by defining a new class of low mobility, static M2M devices which use a modified physical layer control frame structure. The proposal is expected to aid a quick, reliable and cost-effective deployment of M2M devices in the existing GSM networks.
1302.5417
An Ontology Construction Approach for the Domain Of Poultry Science Using Protege
cs.AI
The information retrieval systems that are present nowadays are mainly based on full text matching of keywords or topic based classification. This matching of keywords often returns a large number of irrelevant information and this does not meet the users query requirement. In order to solve this problem and to enhance the search using semantic environment, a technique named ontology is implemented for the field of poultry in this paper. Ontology is an emerging technique in the current field of research in semantic environment. This paper constructs ontology using the tool named Protege version 4.0 and this also generates Resource Description Framework schema and XML scripts for using poultry ontology in web.
1302.5449
Nonparametric Basis Pursuit via Sparse Kernel-based Learning
cs.LG cs.CV cs.IT math.IT stat.ML
Signal processing tasks as fundamental as sampling, reconstruction, minimum mean-square error interpolation and prediction can be viewed under the prism of reproducing kernel Hilbert spaces. Endowing this vantage point with contemporary advances in sparsity-aware modeling and processing, promotes the nonparametric basis pursuit advocated in this paper as the overarching framework for the confluence of kernel-based learning (KBL) approaches leveraging sparse linear regression, nuclear-norm regularization, and dictionary learning. The novel sparse KBL toolbox goes beyond translating sparse parametric approaches to their nonparametric counterparts, to incorporate new possibilities such as multi-kernel selection and matrix smoothing. The impact of sparse KBL to signal processing applications is illustrated through test cases from cognitive radio sensing, microarray data imputation, and network traffic prediction.