1,721,008 research outputs found
Going Beyond Counting First Authors in Author Co-citation Analysis
Full text linkThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
Variations on the Author
Full text link“Variations on the Author” discusses two of Eduardo Coutinho’s recent films (Um Dia na Vida, from 2010, and Últimas Conversas, posthumously released in 2015) and their contribution to the general question of documentary authorship. The director’s filmography is characterized by a consistent yet self-effacing form of authorial self-inscription: Coutinho often features as an interviewer that rather than express opinions propels discourses; an interviewer that is good at listening. This mode of self-inscription characterizes him as an author who is not expressive but who is nonetheless markedly present on the screen. In Um Dia na Vida, however, Coutinho is completely absent form the image, while Últimas Conversas, on the contrary, includes a confessional prologue that moves the director from the margins to the center of his films. This article examines the ways in which these works stand out in the filmography of a director who offers new insights into the notion of cinematic authorship
Appropriate Similarity Measures for Author Cocitation Analysis
Full text linkWe provide a number of new insights into the methodological discussion about author cocitation analysis. We first argue that the use of the Pearson correlation for measuring the similarity between authors’ cocitation profiles is not very satisfactory. We then discuss what kind of similarity measures may be used as an alternative to the Pearson correlation. We consider three similarity measures in particular. One is the well-known cosine. The other two similarity measures have not been used before in the bibliometric literature. Finally, we show by means of an example that our findings have a high practical relevance.information science;Pearson correlation;cosine;similarity measure;author cocitation analysis
Recommended from our members
Approximate Inference in Graphical Models
Full text linkGraphical models have become a central paradigm for knowledge representation and rea- soning over models with large numbers of variables. Any useful application of these models involves inference, or reasoning about the state of the underlying variables and quantify- ing the models’ uncertainty about any assignment to them. Unfortunately, exact inference in graphical models is fundamentally intractable, which has led to significant interest in approximate inference algorithms.In this thesis we address several aspects of approximate inference that affect its quality. First, combining the ideas from variational inference and message passing on graphical models, we study how the regions over which the approximation is formed can be selected more effectively using a content-based scoring function that computes a local measure of the improvement to the upper bound to log partition function. We then extend this framework to use the available memory more efficiently, and show that this leads to better approximations. We propose different memory allocation strategies and empirically show how they can improve the quality of the approximation to the upper bound. Finally, we address the optimization algorithms used in approximate inference tasks. Focusing on maximum a posteriori (MAP) inference and linear programming (LP), we show how the Alternating Direction Method of Multipliers (ADMM) technique can provide an elegant algorithm for finding the saddle point of the augmented Lagrangian of the approximation, and present an ADMM-based algorithm to solve the primal form of the MAP-LP whose closed form updates are based on a linear approximation technique
Variational Methods for Optimal Experimental Design
Full text linkIn this work we study variational methods for Bayesian optimal experimental design (BOED). Experimentation is a cornerstone of science and is central to any major engineering effort. Often experiments require the use of substantial resources, from expensive equipment to limited researcher time; in addition, experiments can be dangerous or may be required to be completed in a given period of time. For these reasons, we prefer to conduct our experiments as efficiently as possible, acquiring as much information as we can given the resources available to us. Optimal experimental design (OED) is a sub-field of statistics focused on developing methods for accomplishing this goal. The OED problem is formulated by defining a utility function over designs and optimizing this function over the set of all feasible designs. We focus on the \emph{Expected Information Gain} (EIG), a widely used utility function with sound theoretical support. However, in practice the EIG is intractable to compute, and approximation strategies are required. We investigate the use of variational methods for this purpose and show substantial improvement over competing approximation techniques. A specific form of OED common in the field of machine learning (ML) is \emph{active learning} (AL). In the active learning framework, we would like to obtain a labeled dataset in order to train a supervised model. However, for all the reasons stated, labeling data points can be costly and again we should make efficient use of our labeling resources. We present a novel application of active learning to optimize spectroscopic follow up for large scale astronomical surveys. Finally, much of this work requires learning functions over sets which we know must satisfy certain properties (e.g., permutation invariance). We conclude the thesis by presenting a novel neural network architecture for predicting the astronomical class of individual objects in the same exposure using a neural architecture specifically designed to accommodate known inductive biases present in the data
Recommended from our members
Probabilistic Models for Brain Image Collection, Classication, and Functional Connectivity.
Full text linkThe use of functional neuroimaging to evaluate brain disorders has become pervasive in the scientific community. The technique provides researchers with a means to evaluate dynamic in-vivo brain function. Over the last thirty years of using neuroimaging techniques to evaluate brain disorders, there is evidence suggesting some illnesses are characterized by differences in regional brain function whereas others by differences in regional connectivity. Disorders with gross anatomical and functional changes such as Alzheimer's disease and traumatic brain injury are often visually discernible in brain scans and differences quantifiable using typical mass univariate analysis techniques. Conversely, disorders with subtle functional changes (e.g. depression) or subtle changes in how the brain communicates (e.g. schizophrenia) are less amiable to existing analysis techniques. Detecting these subtle differences in molecular imaging data, often plagued by noisy measurements from the imaging system, further impedes our ability to gain valuable insights into brain disorders. In this dissertation we use a variety of tools from machine learning and probabilistic modeling to develop new models for decreasing noise in data captured from our imaging systems, improve feature extraction for detecting differences in regional brain function, and evaluate group-based functional connectivity models and their performance in settings with small sample sizes. Each of these models are presented separately with experiments designed to show improvements over existing methodologies and measures of accuracy in both disease classification and recovering gold-standard functional relationships in the brain
Recommended from our members
Anytime Approximate Inference in Graphical Models
Full text linkGraphical models are a powerful framework for modeling interactions within complex systems. Reasoning over graphical models typically involves answering inference queries, such as computing the most likely configuration (maximum a posteriori or MAP) or evaluating the marginals or normalizing constant of a distribution (the partition function); a task called marginal MAP generalizes these two by maximizing over a subset of variables while marginalizing over the rest.Exact computation of these queries is known to be intractable in general, leading to the development of many approximate schemes, the major categories of which are variational methods, search algorithms, and Monte Carlo sampling. Within these, anytime techniques that provide some guarantees on the correct value, and can be improved with more computational effort, are valued for quickly providing users with confidence intervals or certificates of accuracy and allow users to decide the desired balance of quality, time and memory.In this dissertation, we develop a series of approximate inference algorithms for the partition function and marginal MAP with anytime properties by leveraging ideas and techniques from the three inference paradigms, and integrating them to provide hybrid solutions that inherit the strengths of all three approaches. We propose anytime anyspace best-first search algorithms that provide deterministic bounds on the partition function and marginal MAP. These best-first search schemes take advantage of both AND/OR tree search and optimized variational heuristics. We then extend this approach to give anytime probabilistic confidence bounds via a dynamic importance sampling algorithm, which interleaves importance sampling (using proposal distributions extracted from the variational bound) with our best-first search algorithm to refine the proposal. We also propose a framework for interleaving sampling with the optimization of the initial variational bound, which can automatically balance its computational effort between the two schemes. Overall, we show that our hybrid algorithms perform significantly better than existing methods, giving flexible approaches with excellent anytime confidence bounds
Dispelling the Myths Behind First-author Citation Counts
Full text linkWe conducted a full-scale evaluative citation analysis study of scholars in the XML research field to explore just how different from each other author rankings resulting from different citation counting methods actually are, and to demonstrate the capability of emerging data and tools on the Web in supporting more realistic citation counting methods. Our results contest some common arguments for the continued
use of first-author citation counts in the evaluation of scholars, such as high correlations between author rankings by first-author citation counts and other citation
counting methods, and high costs of using more realistic citation counting methods that are not well-supported by the ISI databases. It is argued that increasingly available digital full text research papers make it possible for citation analysis studies to go beyond what the ISI databases have directly supported and to employ more
sophisticated methods
Recommended from our members
Reasoning and Decisions in Probabilistic Graphical Models - A Unified Framework
Full text linkProbabilistic graphical models such as Markov random fields, Bayesian networks and decision networks (a.k.a. influence diagrams) provide powerful frameworks for representing and exploiting dependence structures in complex systems. However, making predictions or decisions using graphical models involve challenging computational problems of optimization and/or estimation in high dimensional spaces. These include combinatorial optimization tasks such as maximum a posteriori (MAP), which finds the most likely configuration, or marginalization tasks that calculate the normalization constants or marginal probabilities. Even more challenging tasks require a hybrid of both: marginal MAP tasks find the optimal MAP prediction while marginalizing over missing information or latent variables, while decision-making problems search for optimal policies over decisions in single- or multi-agent systems, in order to maximize expected utility in uncertain environments.All these problems are generally NP-hard, creating a need for efficient approximations. The last two decades have witnessed significant progress on traditional optimization and marginalization problems, especially via the development of variational message passing algorithms. However, there has been less progress on the more challenging marginal MAP and decision-making problems.This thesis presents a unified variational representation for all these problems. Based on our framework, we derive a class of efficient algorithms that combines the advantages of several existing algorithms, resulting in improved performance on traditional marginalization and optimization tasks. More importantly, our framework allows us to easily extend most or all existing variational algorithms to hybrid inference and decision-making tasks, and significantly improves our ability to solve these difficult problems. In particular, we propose a spectrum of efficient belief propagation style algorithms with "message passing" forms, which are simple, fast and amenable to parallel or distributed computation. We also propose a set of convergent algorithms based on proximal point methods, which have the nice form of transforming the hybrid inference problem into a sequence of standard marginalization problems. We show that our algorithms significantly outperform existing approaches in terms of both empirical performance and theoretical properties
- …
