Universidad del Noreste

Modern machine learning suffers from \textit{catastrophic forgetting} when learning new classes incrementally. The performance dramatically degrades due to the missing data of old classes. Incremental learning methods have been proposed to retain the knowledge acquired from the old classes, by using knowledge distilling and keeping a few exemplars from the old classes. However, these methods struggle to \textbf{scale up to a large number of classes}. We believe this is because of the combination of two factors: (a) the data imbalance between the old and new classes, and (b) the increasing number of visually similar classes. Distinguishing between an increasing number of visually similar classes is particularly challenging, when the training data is unbalanced. We propose a simple and effective method to address this data imbalance issue. We found that the last fully connected layer has a strong bias towards the new classes, and this bias can be corrected by a linear model. With two bias parameters, our method performs remarkably well on two large datasets: ImageNet (1000 classes) and MS-Celeb-1M (10000 classes), outperforming the state-of-the-art algorithms by 11.1\% and 13.2\% respectively.

The authors propose a theory of economic development in which technology adoption and barriers to such adoptions are the focus. The size of these barriers differs across countries and time. The larger these barriers, the greater the investment a firm must make to adopt a more advanced technology. The model is calibrated to the U.S. balanced growth observations and the postwar Japanese development miracle. For this calibrated structure, the authors find that the disparity in technology adoption barriers needed to account for the huge observed income disparity across countries is not implausibly large. Copyright 1994 by University of Chicago Press.

The ability of the emotion gratitude to shape costly prosocial behavior was examined in three studies employing interpersonal emotion inductions and requests for assistance. Study 1 demonstrated that gratitude increases efforts to assist a benefactor even when such efforts are costly (i.e., hedonically negative), and that this increase differs from the effects of a general positive affective state. Additionally, mediational analyses revealed that gratitude, as opposed to simple awareness of reciprocity norms, drove helping behavior. Furthering the theory that gratitude mediates prosocial behavior, Study 2 replicated the findings of Study 1 and demonstrated gratitude's ability to function as an incidental emotion by showing it can increase assistance provided to strangers. Study 3 revealed that this incidental effect dissipates if one is made aware of the true cause of the emotional state. Implications of these findings for the role of gratitude in building relationships are discussed.

Much of the stress and strain of student life remains hidden. The StudentLife continuous sensing app assesses the day-to-day and week-by-week impact of workload on stress, sleep, activity, mood, sociability, mental well-being and academic performance of a single class of 48 students across a 10 week term at Dartmouth College using Android phones. Results from the StudentLife study show a number of significant correlations between the automatic objective sensor data from smartphones and mental health and educational outcomes of the student body. We also identify a Dartmouth term lifecycle in the data that shows students start the term with high positive affect and conversation levels, low stress, and healthy sleep and daily activity patterns. As the term progresses and the workload increases, stress appreciably rises while positive affect, sleep, conversation and activity drops off. The StudentLife dataset is publicly available on the web.

Abstract We outline the need to, and provide a guide on how to, conduct a meta‐analysis on one's own studies within a manuscript. Although conducting a “mini meta” within one's manuscript has been argued for in the past, this practice is still relatively rare and adoption is slow. We believe two deterrents are responsible. First, researchers may not think that it is legitimate to do a meta‐analysis on a small number of studies. Second, researchers may think a meta‐analysis is too complicated to do without expert knowledge or guidance. We dispel these two misconceptions by (1) offering arguments on why researchers should be encouraged to do mini metas, (2) citing previous articles that have conducted such analyses to good effect, and (3) providing a user‐friendly guide on calculating some meta‐analytic procedures that are appropriate when there are only a few studies. We provide formulas for calculating effect sizes and converting effect sizes from one metric to another (e.g., from Cohen's d to r ), as well as annotated Excel spreadsheets and a step‐by‐step guide on how to conduct a simple meta‐analysis. A series of related studies can be strengthened and better understood if accompanied by a mini meta‐analysis.

Most recent semantic segmentation methods adopt a U-Net framework with an encoder-decoder architecture. It is still challenging for U-Net with a simple skip connection scheme to model the global multi-scale context: 1) Not each skip connection setting is effective due to the issue of incompatible feature sets of encoder and decoder stage, even some skip connection negatively influence the segmentation performance; 2) The original U-Net is worse than the one without any skip connection on some datasets. Based on our findings, we propose a new segmentation framework, named UCTransNet (with a proposed CTrans module in U-Net), from the channel perspective with attention mechanism. Specifically, the CTrans (Channel Transformer) module is an alternate of the U-Net skip connections, which consists of a sub-module to conduct the multi-scale Channel Cross fusion with Transformer (named CCT) and a sub-module Channel-wise Cross-Attention (named CCA) to guide the fused multi-scale channel-wise information to effectively connect to the decoder features for eliminating the ambiguity. Hence, the proposed connection consisting of the CCT and CCA is able to replace the original skip connection to solve the semantic gaps for an accurate automatic medical image segmentation. The experimental results suggest that our UCTransNet produces more precise segmentation performance and achieves consistent improvements over the state-of-the-art for semantic segmentation across different datasets and conventional architectures involving transformer or U-shaped framework. Code: https://github.com/McGregorWwww/UCTransNet.

Both pattern recognition and computer vision have experienced rapid progress in the last twenty-five years. This book provides the latest advances on pattern recognition and computer vision along with their many applications. It features articles written by renowned leaders in the field while topics are presented in readable form to a wide range of readers. The book is divided into five parts: basic methods in pattern recognition, basic methods in computer vision and image processing, recognition applications, life science and human identification, and systems and technology.

This article examines conceptual issues relating to continuity and change in crime over the life course. Building on past efforts, we first distinguish self‐selection from a cumulative, developmental process whereby delinquent behavior attenuates adult social bonds (e.g., labor force attachment, marital cohesion). We then conceptualize various types of change and argue that social capital and turning points are crucial in understanding processes of change in the adult life course. These concepts are illustrated by examining person‐based, life‐history data drawn from the Gluecks' longitudinal study of 1,000 men. Although adult crime is clearly connected to childhood behavior, these qualitative data suggest that both incremental and abrupt change are structured by changes in adult social bonds. We conclude with some hypotheses and implications for future research on subjective contingencies, opportunity structures, and chance encounters as potential turning points for change, especially as they interact with race, class location, and historical context.

Causal complexity has long been recognized as a ubiquitous feature underlying organizational phenomena, yet current theories and methodologies in management are for the most part not well-suited to its direct study. The introduction of the Qualitative Comparative Analysis (QCA) configurational approach has led to a reinvigoration of configurational theory that embraces causal complexity explicitly. We argue that the burgeoning research using QCA represents more than a novel methodology; it constitutes the emergence of a neo-configurational perspective to the study of management and organizations that enables a fine-grained conceptualization and empirical investigation of causal complexity through the logic of set theory. In this article, we identify four foundational elements that characterize this emerging neo-configurational perspective: (a) conceptualizing cases as set theoretic configurations, (b) calibrating cases’ memberships into sets, (c) viewing causality in terms of necessity and sufficiency relations between sets, and (d) conducting counterfactual analysis of unobserved configurations. We then present a comprehensive review of the use of QCA in management studies that aims to capture the evolution of the neo-configurational perspective among management scholars. We close with a discussion of a research agenda that can further this neo-configurational approach and thereby shift the attention of management research away from a focus on net effects and towards examining causal complexity.

BACKGROUND: Hospitalization and subsequent discharge home often involve discontinuity of care, multiple changes in medication regimens, and inadequate patient education, which can lead to adverse drug events (ADEs) and avoidable health care utilization. Our objectives were to identify drug-related problems during and after hospitalization and to determine the effect of patient counseling and follow-up by pharmacists on preventable ADEs. METHODS: We conducted a randomized trial of 178 patients being discharged home from the general medicine service at a large teaching hospital. Patients in the intervention group received pharmacist counseling at discharge and a follow-up telephone call 3 to 5 days later. Interventions focused on clarifying medication regimens; reviewing indications, directions, and potential side effects of medications; screening for barriers to adherence and early side effects; and providing patient counseling and/or physician feedback when appropriate. The primary outcome was rate of preventable ADEs. RESULTS: Pharmacists observed the following drug-related problems in the intervention group: unexplained discrepancies between patients' preadmission medication regimens and discharge medication orders in 49% of patients, unexplained discrepancies between discharge medication lists and postdischarge regimens in 29% of patients, and medication nonadherence in 23%. Comparing trial outcomes 30 days after discharge, preventable ADEs were detected in 11% of patients in the control group and 1% of patients in the intervention group (P = .01). No differences were found between groups in total ADEs or total health care utilization. CONCLUSIONS: Pharmacist medication review, patient counseling, and telephone follow-up were associated with a lower rate of preventable ADEs 30 days after hospital discharge. Medication discrepancies before and after discharge were common targets of intervention.

Prior research on victimization in the United States has generally neglected two key areas—victimization among juveniles and young adults and the connection between offending and victimization. The research presented here fuses these two concerns by examining the effect of delinquent lifestyles on the criminal victimization of teenagers and young adults. An examination of longitudinal data from the first five waves of the National Youth Survey suggests that adolescent involvement in delinquent lifestyles strongly increases the risk of both personal and property victimization. Further, the analysis reveals that a significant proportion of the risk of victimization incurred by different demographic subgroups—especially males—results from greater involvement in lifestyles characterized by delinquency. The authors conclude that victimization patterns among youths cannot be understood apart from criminal and deviant activities.

Every second, the thoughts and feelings of millions of people across the world are recorded in the form of 140-character tweets using Twitter. However, despite the enormous potential presented by this remarkable data source, we still do not have an understanding of the Twitter population itself: Who are the Twitter users? How representative of the overall population are they? In this paper, we take the first steps towards answering these questions by analyzing data on a set of Twitter users representing over 1% of the U.S. population. We develop techniques that allow us to compare the Twitter population to the U.S. population along three axes (geography, gender, and race/ethnicity), and find that the Twitter population is a highly non-uniform sample of the population.

Management scholars study phenomena marked by complex interdependencies where multiple explanatory factors combine to bring about an outcome of interest. Yet, theorizing about causal complexity can prove challenging for the correlational theorizing that is predominant in the field of management, given its “net effects thinking” that emphasizes the unique contribution of individual explanatory factors. In contrast, configurational theories and thinking are well-suited to explaining causally complex phenomena. In this article, we seek to advance configurational theorizing by providing a model of the configurational theorizing process which consists of three iterative stages—scoping, linking and naming. In each stage, we develop and offer several heuristics aimed at stimulating configurational theorizing. That is, these theorizing heuristics are intended to help scholars discover configurations of explanatory factors, probe the connections among these factors, and articulate the orchestrating themes that underpin their coherence. We conclude with a discussion of how configurational theorizing advances theory development in the field of management and organizations, and beyond.

Receptors for ATP and ADP and adenosine exert various effects. ATP and ADP signaling is mainly proinflammatory, and adenosine signaling is mainly antiinflammatory. Receptors for these nucleosides are emerging as therapeutic targets in a number of inflammatory and autoimmune diseases.

Two head structures (i.e. fully connected head and convolution head) have been widely used in R-CNN based detectors for classification and localization tasks. However, there is a lack of understanding of how does these two head structures work for these two tasks. To address this issue, we perform a thorough analysis and find an interesting fact that the two head structures have opposite preferences towards the two tasks. Specifically, the fully connected head (fc-head) is more suitable for the classification task, while the convolution head (conv-head) is more suitable for the localization task. Furthermore, we examine the output feature maps of both heads and find that fc-head has more spatial sensitivity than conv-head. Thus, fc-head has more capability to distinguish a complete object from part of an object, but is not robust to regress the whole object. Based upon these findings, we propose a Double-Head method, which has a fully connected head focusing on classification and a convolution head for bounding box regression. Without bells and whistles, our method gains +3.5 and +2.8 AP on MS COCO dataset from Feature Pyramid Network (FPN) baselines with ResNet-50 and ResNet-101 backbones, respectively.

Abstract. This study shows that financial analysts of brokerage firms that provide investment banking services to a company (investment banker analysts) are optimistic, relative to other (noninvestment banker) analysts, in their earnings forecasts and investment recommendations. Returns earned by following the investment recommendations of investment banker analysts, however, are not significantly different from those of non‐investment banker analysts. Given that information regarding the investment banking relationships of brokerage firms is publicly available, we find evidence that capital market participants rely relatively less on the investment banker analysts in forming their earnings expectations. Although we find a significant capital market reaction around the noninvestment banker analysts' research report dates and not around the investment banker analysts' research report dates, the difference between the two market reactions is not statistically significant. Finally, we find that investment banker analysts' earnings forecasts are, on average, as accurate as those of noninvestment banker analysts. Résumé. Les auteurs mettent en évidence le fait que les analystes financiers des maisons de courtage qui offrent des services de prise ferme aux entreprises (les analystes de courtiers preneurs ferme) sont optimistes dans leurs prévisions de bénéfices et leurs recommandations de placements, par comparaison aux autres analystes (c'est‐à‐dire aux analystes de courtiers qui ne sont pas preneurs ferme). Les rendements obtenus par les investisseurs qui observent les recommandations de placements des analystes de courtiers preneurs ferme ne sont cependant pas sensiblement différents de ceux qu'obtiennent les investisseurs qui se fient aux analystes des courtiers qui ne sont pas preneurs ferme. Compte tenu du fait que l'information relative aux relations qu'entretiennent les maisons de courtage en matière de prise ferme est du domaine public, les constatations des auteurs confirment que les participants au marché financier s'appuient relativement moins sur le verdict des analystes des courtiers preneurs ferme dans le calcul de leur espérance de gains. Bien que les auteurs observent une réaction marquée du marché financier à proximité des dates de publication des rapports de recherche des analystes des courtiers qui ne sont pas preneurs ferme, ce qui n'est pas le cas à proximité des dates de publication des rapports de recherche des analystes des courtiers preneurs ferme, la différence entre ces deux réactions n'est pas statistiquement significative. Enfin, les auteurs constatent que les prévisions de bénéfices des analystes des courtiers preneurs ferme sont, en moyenne, aussi exactes que celles des analystes des courtiers qui ne sont pas preneurs ferme.

ADVERTISEMENT RETURN TO ISSUEPREVArticleNEXTFerromagnetic molecular charge-transfer complexesJoel S. Miller, Arthur J. Epstein, and William M. ReiffCite this: Chem. Rev. 1988, 88, 1, 201–220Publication Date (Print):January 1, 1988Publication History Published online1 May 2002Published inissue 1 January 1988https://pubs.acs.org/doi/10.1021/cr00083a010https://doi.org/10.1021/cr00083a010research-articleACS PublicationsRequest reuse permissionsArticle Views2413Altmetric-Citations591LEARN ABOUT THESE METRICSArticle Views are the COUNTER-compliant sum of full text article downloads since November 2008 (both PDF and HTML) across all institutions and individuals. These metrics are regularly updated to reflect usage leading up to the last few days.Citations are the number of other articles citing this article, calculated by Crossref and updated daily. Find more information about Crossref citation counts.The Altmetric Attention Score is a quantitative measure of the attention that a research article has received online. Clicking on the donut icon will load a page at altmetric.com with additional details about the score and the social media presence for the given article. Find more information on the Altmetric Attention Score and how the score is calculated. Share Add toView InAdd Full Text with ReferenceAdd Description ExportRISCitationCitation and abstractCitation and referencesMore Options Share onFacebookTwitterWechatLinked InRedditEmail Other access optionsGet e-Alertsclose Get e-Alerts

The mainstream paradigm behind continual learning has been to adapt the model parameters to non-stationary data distributions, where catastrophic forgetting is the central challenge. Typical methods rely on a rehearsal buffer or known task identity at test time to retrieve learned knowl-edge and address forgetting, while this work presents a new paradigm for continual learning that aims to train a more succinct memory system without accessing task identity at test time. Our method learns to dynamically prompt (L2P) a pre-trained model to learn tasks sequen-tially under different task transitions. In our proposed framework, prompts are small learnable parameters, which are maintained in a memory space. The objective is to optimize prompts to instruct the model prediction and ex-plicitly manage task-invariant and task-specific knowledge while maintaining model plasticity. We conduct comprehen-sive experiments under popular image classification bench-marks with different challenging continual learning set-tings, where L2P consistently outperforms prior state-of-the-art methods. Surprisingly, L2P achieves competitive results against rehearsal-based methods even without a re-hearsal buffer and is directly applicable to challenging task-agnostic continual learning. Source code is available at https://github.com/google-research/12p.

Abstract There has been growing recognition in recent years of the importance of corporate governance in ensuring sound financial reporting and deterring fraud. The audit serves as a monitoring device and is thus part of the corporate governance mosaic. The objective of this paper is to examine the impact of various corporate governance factors, such as the board of directors and the audit committee, on the audit process. Importantly, there is little professional guidance on how auditors should consider such factors when formulating an appropriate audit strategy, and there has been only one prior study on this issue (Cohen and Hanno 2000). Because there are no current specific auditing standards that relate to the effect of corporate governance on the audit process, we conducted a semi‐structured interview with 36 auditors on current audit practices in considering corporate governance in the audit process. Reflecting on client experiences, auditors indicate a range of views with regard to the elements included in the rubric of “corporate governance”. Most significantly, auditors view management as the primary driver of corporate governance. The inclusion of top management in the “corporate governance mosaic” is inconsistent with agency theory's prescription of the board and other mechanisms serving as a means to independently oversee management's actions to protect stakeholders. Auditors consider corporate governance factors to be especially important in the client acceptance phase and in an international context. Further, despite the attention placed on the audit committee in the academic literature, in the business community, and by regulators in different countries (e.g., Canada, United States, Australia), several respondents indicated that their experiences with their clients suggest that audit committees are typically ineffective and lack sufficient power to be a strong governance mechanism. Implications for research and practice are presented.